Spatial constraints and cell surface molecule depletion structure a randomly connected learning circuit

Emma M Thornton-Kolbe; Maria Ahmed; Finley R Gordon; Bogdan Sieriebriennikov; Donnell L Williams; Yerbol Z Kurmangaliyev; E Josephine Clowney

doi:10.1016/j.cub.2025.05.062

. Author manuscript; available in PMC: 2025 Dec 20.

Published in final edited form as: Curr Biol. 2025 Jun 26;35(13):3191–3208.e10. doi: 10.1016/j.cub.2025.05.062

Spatial constraints and cell surface molecule depletion structure a randomly connected learning circuit

Emma M Thornton-Kolbe ^1,^*, Maria Ahmed ^2,^$,^*, Finley R Gordon ², Bogdan Sieriebriennikov ³, Donnell L Williams ², Yerbol Z Kurmangaliyev ⁴, E Josephine Clowney ^2,^5,^6,^#

PMCID: PMC12716490 NIHMSID: NIHMS2125319 PMID: 40578357

Summary

The brain can represent almost limitless objects to “categorize an unlabeled world” (Edelman, 1989). This feat is supported by expansion layer circuit architectures, in which neurons carrying information about discrete sensory channels make combinatorial connections onto much larger postsynaptic populations. Combinatorial connections in expansion layers are modeled as randomized sets. The extent to which randomized wiring exists in vivo is debated, and how combinatorial connectivity patterns are generated during development is not understood. Non-deterministic wiring algorithms could program such connectivity using minimal genomic information. Here, we investigate anatomic and transcriptional patterns and perturb partner availability to ask how Kenyon cells, the expansion layer neurons of the insect mushroom body, obtain combinatorial input from olfactory projection neurons. Olfactory projection neurons form their presynaptic outputs in an orderly, predictable, and biased fashion. We find that Kenyon cells accept spatially co-located but molecularly heterogeneous inputs from this orderly map, and ask how their cell surface molecule expression impacts partner choice. Cell surface immunoglobulins are broadly depleted in Kenyon cells, and we propose that this allows them to form connections with molecularly heterogeneous partners. This model can explain how developmentally identical neurons acquire diverse wiring identities.

Introduction

In expansion layer circuits, large sets of developmentally homogenous neurons receive combinatorial sensory inputs from more diverse presynaptic partners.^1-7 This expands the diversity of perceivable stimuli from the raw number of sensory channels to a much larger set of combinations.^8-11 Expansion layers seem to have evolved multiple times across Bilaterian clades and include the parallel lobe system of cephalopods, the arthropod mushroom body, and the vertebrate pallium, hippocampus, and cerebellum.^12-16 These circuit architectures are often used for associative learning, in which previously arbitrary sensory representations become connected with temporally linked events.

The density and identity of inputs to expansion layer neurons are set up during developmental wiring and bound the functionality of the circuit. The Drosophila melanogaster mushroom body provides an excellent model to explore mechanisms that allow expansion layer neurons to receive diverse sets of presynaptic inputs. Here, ~2000 Kenyon cells per hemisphere each receive 3-10 inputs from among 51 olfactory channels.^11,17-19 Kenyon cells dictate the density of these inputs during development and the diversity of their odor responses diminishes if input density is too low or too high.^20,21 Different cells are sensitive to different stimuli because they receive different input combinations.^6,18,22-25 The developmental mechanisms that allow this diversity of inputs are unknown and are unlikely to mimic those observed in well-studied motor or sensory structures, where neurons with diverse developmental identities connect to one another in precise ways.

By contrast, the developmental identities of expansion layer neurons appear insufficient to predict their wiring identities: Kenyon cells (KCs) are developmentally homogenous; they are born from four neural stem cells (neuroblasts) and comprise only seven major anatomic subtypes in the adult.^17,26,27 KC types are born sequentially, and each of the four neuroblasts makes all types in parallel.²⁸ Presynaptic inputs to KCs are much more diverse: 51 types of olfactory projection neurons (PNs) arise in a precise and reproducible order from two distinct neuroblasts and each carry information to the mushroom body from one glomerulus in the antennal lobe.^29,30

Here, we ask how KCs with shared specification programs can nevertheless acquire distinct combinations of presynaptic inputs during development. There has been enduring debate in the field about the extent to which KC inputs are random.^{6,18,22-25,31-40} and a similar debate emerges in the cerebellum literature.^5,7,41-45 We use electron microscopy data⁴⁶ to unify previous reports^{26,28,32-34,39,47,48} and find that PNs of different types innervate unique and reproducible regions within the mushroom body calyx.^{25,34,39,48,49} We find that individual KCs are spatially constrained depending on their neuroblast origin and subtype, and that these spatial constraints, overlaid on the orderly PN array, set up predictable biases in the boutons that individual cells can reach. Within domains, KCs sample randomly from the available PN distribution.

Different PN types express distinct repertoires of cell surface molecules.^50,51 Therefore, we next characterize the transcriptional programs of developing KCs to ask how neurons of homogenous developmental identity can synapse with molecularly diverse partners. We find that KCs have a transcriptional depletion of immunoglobulin superfamily proteins. Genes in this family are proposed to regulate a variety of neuron-type-specific developmental processes, including synaptic partner choice.^52-56 We propose that downregulation of immunoglobulins in KCs blinds them to the molecular differences among potential partners. Each KC is then free to obtain inputs promiscuously from among the PN boutons available within its domain, generating diverse wiring inputs to developmentally homogenous cells.

Results

Projection neuron anatomy in the mushroom body calyx is orderly, biased, and predictable

Uniglomerular, cholinergic olfactory PNs are produced by invariant neural lineages and are predictable in the antennal lobe glomeruli they innervate and in the number of boutons they elaborate in the calyx.^{19,29,30,32,33,57} Furthermore, many early light microscopy studies identified clear and predictable differences in the positioning of neurites and boutons of different PN types.^32-34,39,48 Using the Hemibrain connectome, we sought (1) to situate the anatomies of individual PN types within a holistic description of bouton placement and (2) to identify developmental variables producing these anatomies in order to build a framework that reconciles predictable PN development with variation in KC connectivity.

In this connectome, 105 PNs provide input to 1728 olfactory KCs via their presynapse-dense terminal boutons in the calyx (Figure 1A-C).^58-61 An additional ~200 KCs of the αβ-p and γ-d types receive non-olfactory, mostly visual input.³¹ We identified 24,791 synapses from olfactory PNs to olfactory KCs and clustered these synapses into boutons (see Methods). To simplify further analyses, we rotated coordinates, which are tilted relative to the whole brain, to align with Cartesian axes (Figure S1A-B, Methods).

First we looked at the positioning of boutons across the 20μm anterior-posterior (A-P) span of the calyx (Figure 1D). The PN axons in the medial antennal lobe tract (mALT) traverse the ventral side of the calyx, and individual axons are segregated in the mALT based on the developmental stage of their birth and neuroblast origin.⁴⁸ We found the A-P position of boutons is well predicted by the position of the axon from which they arise (Figure 1E). Thus, bouton position reflects developmental history: axons and boutons of neurons born embryonically are broadly distributed along the A-P axis, while axons and boutons of larvally-born PNs from the anterodorsal (adPN) and lateral (lPN) lineages are biased toward anterior versus posterior domains, respectively (Figure S1C,D,E).

We next looked at positioning of boutons along the 50μm medial-lateral (M-L) axis of the calyx (Figure 1F). While some boutons form directly on the axon, most form at the end of collaterals. The M-L position of a bouton is strongly predicted by the M-L position of the branch point of its collateral from the main axon (Figure 1G). PNs in the Hemibrain have 1-7 collaterals; we found that most neurons with just one or two placed them toward the center of the calyx while those with 3 or more spread them further along the M-L axis (Figure 1H,I). These highly branched PNs populating the periphery of the Hemibrain calyx are the same PN types observed to populate this domain in other studies (Figure S1F,G).^25,34

Finally, we described the positioning of boutons along the 30μm dorsal-ventral (D-V) axis (Figure 1F). While boutons are most densely packed in the ventral calyx closer to the axons, some are placed more dorsally at the end of long collaterals. Indeed, collateral length predicts the D-V positioning of the bouton and varies equally across both other axes (Figure 1J, H).

While the Hemibrain represents just a single brain instance, 29 of 51 glomeruli are innervated by multiple sister PNs. This cell type redundancy allowed us to ask how closely neurons of the same type mirrored each other. We measured the distance between cognate boutons (those with the same M-L ordinal position) on neurons of the same and different PN types (Figure 1K). We found that cognate boutons on PNs of the same type were closer together (Figure 1L). Bouton position is the result of collateral placement and length. While collateral number is predicted by PN type and impacts M-L collateral spread, as above, fine branch point positions along the M-L axis were not type-specific (Figure 1M). In contrast, cognate collaterals on PNs of the same type tend to be more similar in length than cognate collaterals on PNs of different types (Figure 1N). Type-specific collateral length and placement along the axon could result from cell-autonomous differences in neurite behavior or from reproducible temporal stratification of PN incursions into the calyx. Similarity in axonal arbor shape across sister PNs of the same type was also observed in the Female Adult Fly Brain (FAFB) dataset⁴⁹ and we observe remarkable concordance of PN type behavior between the Hemibrain dataset and many previous light and EM analyses. This demonstrates a high degree of consistency and predictability in bouton placement in the calyx, such that boutons carrying information about different odors are distributed in a mixed but developmentally non-random manner across brains (Figure S1G).^{25,32-34,39,48}

Individual Kenyon cells innervate restricted regions of the calyx

We next sought to describe the calyx innervation patterns of KCs and the PN repertoires that each can access. Four mushroom body neuroblasts per hemisphere produce each of the KC types sequentially: γ KCs are born during the embryonic and early larval stages, α’β’ during late larval stages, and αβ in pupae (Figure 2A).^27,28 To examine how these developmental traits influence calyx organization, we ascertained the clonal origin of each KC by measuring its axonal position outside the calyx, as it traverses the pedunculus on its way to the lobes (see Methods); for γ, α’β’, and αβ type assignment, we used Hemibrain annotations.^23,46

Figure 2: — **(A)** Numbers of olfactory KCs assigned to each neuroblast and type in the Hemibrain (top). Calyx anatomy of a KC (bottom).

**(B-C)** Centroids of KC claws colored by predicted neuroblast origin (B) or annotated KC type (C). Below: Histograms of claw center locations.

**(D)** Distribution of input from each PN type to each KC group.

**(E)** Hemibrain skeletons of a DC1 PN (top) and a DM3 PN (bottom) Dotted line: calyx.

**(F)** Fractions of KCs postsynaptic to boutons labeled in (E) by clonal origin and type.

**(G)** PN bouton M-L position predicts clonal origin of KC partners (left) and A-P position predicts partner KC type (right). Pink and cyan dots are boutons labeled in (E).

**(H)** Distributions of the spread of claws of each KC. Individual KCs do not spread their claws along the whole M-L axis, but do reach across A-P and D-V axes. All graphs show same absolute spatial scale, with percent of that dimension below.

See also Figure S2.

Individual KCs extend 3-10 dendritic claws from the primary neurite within the calyx, each of which enwraps a single PN bouton.^18,59 We found that claws from KCs sharing a neuroblast origin are clustered together along the M-L axis; these clonal claw clouds overlap such there are no sharp boundaries (Figure 2B).^26,39,47 Claws from KCs of each of the three types are found throughout the calyx, though γ claws are enriched in the more anterior part of the calyx, and those of later-born α’β’ and αβ cells in the posterior (Figure 2C).

Our PN analysis predicts that different domains of the calyx will contain regionally unique mixtures of PN bouton types. To test whether KCs inherit these distributions, we described the percent of claws of each clonal or type group receiving input from each PN type. The proportion of overall input contributed by each PN type was heterogeneous and influenced by both the number of boutons of that type and the number of claws enwrapping them (Figure 2D). The average bouton is enwrapped by 22 claws, though the exact number of claws postsynaptic to a bouton is correlated with the number of presynapses it contains, which varies across the boutons of individual PNs, as well as across all anatomical axes of the calyx (Figure S2 A-C). Notably, all PN types connect to KCs born from all four neuroblast clones in the Hemibrain (with the exception of the DL3 PN to clone D KCs) and to all KC types. However, consistent with light level studies of neurite overlap and our analysis of the distribution of PN bouton types (Figure 1), KCs of different clonal origins and of different types received inputs from PN types at different rates.³⁹

To disentangle spatiotemporal from molecular influences on connectivity, we compared the distributions of KCs postsynaptic to boutons that shared molecular similarities (from the same neuron) or spatial similarities (neighbors from different neurons). KC partners of nearby boutons from different PNs were more similar than for distant boutons of the same PN (Figure 2E,F). Across the calyx, the M-L location of a bouton predicts the percent of its postsynaptic partners from each of the four neuroblast clones, while the A-P position predicts the percent of its postsynaptic partners from each of the three types (Figure 2G).

In summary, KCs make dendritic claws whose placement is dependent on their neuroblast origin and cell type. Because they are limited to a region of the calyx, they receive input from PN types at different rates. Connectivity between PNs and KCs is better predicted by the proximity of their neurites than by their cell types. KCs are particularly restricted along the long M-L axis of the calyx, as dendrites of individual cells only reach a quarter of its extent (Figure 2H).

Within a Kenyon cell’s restricted region of the calyx, its inputs are randomized

Previous studies debated the extent to which PN inputs to Kenyon cells are random or structured.^{6,18,22-25,31-40} Combining these reports with our analyses of bouton and claw placement, we hypothesized that spatially restricted KCs use a stochastic developmental mechanism to acquire their inputs from the structured array of PN boutons. To test this, we generated models of calyx connectivity, starting with a model similar to the “random bouton model” in Zheng et al.²⁵ We set PN bouton distribution and KC claw numbers to match the Hemibrain and first allowed simulated KCs to choose boutons randomly. To account for the effect of KC spatial restriction, we then adjusted our model to separately consider KCs from each of the four clones (M-L spatial domains); each of the three types (A-P spatial domains); and the intersection of clonal origins and cell type (12 sub-domains). Here, model KCs could choose only from boutons actually reachable by cells of their type and/or clone in the Hemibrain data (Figure 3A).

Figure 3: — **(A)** Left: Claw centroids of Hemibrain KC groups in the calyx (dotted outline). Right: Centroids of PN boutons connected to those KCs shown at left.

**(B)** Conditional input analysis z-scores for all pairs of PN types for example groups of KCs. Black bar: median. Z-scores >5 or <−5 not shown. This excludes the highest number of z-scores from the “all KCs” model where 151 (5.8%) are excluded.

**(C-D)** Number of KCs connected to both VA2 and DP1l (C) or VA2 and DA2 (D) PNs in the Hemibrain data (red dot) and predicted by 10,000 random models (grey bars).

**(E)** Calyx Hemibrain skeletons of VA2 with DP1l or DA2.

**(F)** Pairwise distances between boutons that share 0 or at least 1 KC partner along each axis. Wilcoxon rank sum test.

For each model type, we generated 10,000 random connectivity matrices. We compared these to the connectivity observed in the Hemibrain using conditional input analysis.²⁵ Observed connectivity in the Hemibrain calyx statistically diverged from that predicted by the full calyx random models. The pairs of PN types that most diverged in the observed data were similar to a set of PN types reported as “overconvergent” in analysis of different brain, the FAFB volume.²⁵ These PNs are also the types we describe in Figure 1H as having higher numbers of collaterals positioned more peripherally and match the “outer” group of PNs in Tanaka et al.³⁴

As we moved from modeling the whole Hemibrain calyx to KCs in developmentally restricted subdomains, observed connectivity more closely recapitulated predictions from random models (Figure 3B). To illustrate the different levels of accuracy among models, we picked the pair of PN types most over-convergent in the Hemibrain calyx compared to the full calyx models: VA2 and DP1l (Figure 3C,E). The whole calyx random model predicts an average of 15 KCs that receive input from these two glomeruli, but 61 Hemibrain KCs actually do. On the other hand, the model predicts DA2 to be co-sampled ~30 times with VA2, but only 9 Hemibrain KCs actually receive both inputs (Figure 3D,E). While the observed co-samplings of these pairs differ dramatically from the prediction based on the full calyx models, they differ minimally from our domain-based model anchored in development.

Finally, we found that pairs of boutons co-innervating at least one KC were closer on the M-L axis than boutons not sharing KCs, but had minimal unique proximity along the D-V or A-P axes (Figure 3F). This suggests that the spatially constrained dendritic development of KCs from different neuroblast clones is the primary determinant of correlation structure in PN inputs to individual cells.

Kenyon cells express a reduced set of cell surface molecules

Our anatomical observations appear to be in conflict: From the PN point of view, the calyx is stereotyped and orderly. On the other hand, while KCs have their own spatial constraints, they don’t appear to discriminate among the PNs they encounter. Different PN types express numerous and diverse cell surface molecules.^50,51 To understand the cell surface molecules KCs might use to acquire varied PN inputs, we analyzed KC transcription across developmental time (Figure 4A). First, we performed RNA sequencing (RNA-seq) on bulk fluorescence-activated cell-sorted (FAC-sorted)PNs, KCs, and “double negative” cells (neither PNs nor KCs) at 45-46h after puparium formation (APF) and in adulthood (Figure 4B). We characterized the expression of 932 putative cell surface and secreted (CSS) molecules compiled by the Zinn lab.⁶² In PNs, similar numbers of CSS’s were up- and down-regulated compared to double negative cells at both time points, as expected (Figure 4B). In contrast, most CSS’s were downregulated in KCs (Figure 4B). This depletion could reflect either global reduction of these transcripts or selective expression of transcripts in KC subsets.

Figure 4: — **(A)** KCs (cyan) and PNs (magenta) innervating the calyx at 36hAPF, 48hAPF, 60hAPF, and 2-3 day old adults.

**(B)** Expression of 932 cell surface and secreted (CSS) molecules in pooled 45h APF and adult KCs or PNs plotted against expression in unlabeled (non-PN, non-KC) brain cells. Lines represent linear regression of KC (cyan) or PN (pink) data.

**(C-F)** UMAPs of scRNAseq dataset. (C) shows all cells by library (time point, KC driver line). (D) by adult-assigned KC type. (E) subsetted and re-clustered KCs. (F) UMAP in (E) split by stage.

**(G)** Number of KCs per class per stage.

**(H)** Distribution of stages assigned to each KC type.

**(I-J)** Log-normalized, average expression of known KC type marker genes (I) and pan-KC but stage-specific cell adhesion molecules (J) at the different time points. Lines show normalized average expression of that gene in that KC class at the 4 time points.

See also Figures S4.

If individual KCs express stochastic combinations of cell surface molecules, they could actively choose varied PN inputs. If individual KCs weakly express cell surface molecules, they might acquire inputs promiscuously because they lack the molecular tools to differentiate them. We therefore performed single-cell RNA sequencing (scRNAseq) on KCs at 48h APF, when PN:KC contacts first appear; 60h APF, during a wave of synaptogenesis across the central brain ⁶³; and 36h APF, to capture any important transcriptional changes that may occur prior to initial contacts. We also sequenced adult KCs in which the seven developmental types were labeled a priori (see Methods).

Transcriptomic clusters at each pupal time point aligned with the cell types assigned in the adult data and expressed known and novel type-specific markers (Figure 4I).^64-66 Pupal KCs clustered into 6 major groups, listed in order of their birth: γd, γm, α’β’m, α’β’ap, αβp, αβ (Figure 4C-H).^17,26 Pupal αβ cells did not cluster neatly onto earlier-born “shell” and later-born “core” subtypes, likely because αβ core cells aren’t born until late pupal stages. γd and αβp visual KCs^{17,23,31,67,68} clustered separately from olfactory γ and αβ cells at all time points (Figure 4D-F). We cross-checked expression of a subset of CSS genes, DIPs and dprs, across bulk RNAseq, scRNAseq, and using knock-in reporter alleles; all three matched well (Figure S3).

Analysis of transcriptional consistency within canonical Kenyon cell classes

We sought to ascertain whether there was transcriptional heterogeneity within the canonical KC classes that might explain variable connectivity. To do so, we used principal component analysis to examine residual variation within type and stage clusters. These analyses suggested that any remaining variation was technical rather than biological. For example, among γm KCs at 36h APF (Figure S4), PC1 captured transcriptional variation in genes involved in the global maturation of neurons.^69-71 PC1 was likely driven by the differences in the developmental stages of individual animals, as each sample was collected across a four hour window. PC1 also separated two experimental batches with distinct genetic backgrounds. Another set of genes showed dynamic expression across time but was universal across subtypes, including the Immunoglobulin superfamily (Ig-SF) molecules kirrel, beat-VII, and DIP-kappa (Figure 4J). Such genes could allow KCs to form synapses with PNs generally, but are unlikely to diversify connectivity. Expression of some CSS genes correlated with the maturation axis at 36h APF, and we cannot rule out a modest influence of temporal variation, whether in cellular maturation rate or birth order, on connectivity.

Immunoglobulin superfamily genes are depleted in individual Kenyon cells

To test if the bulk depletion of CSS transcripts in KCs (Figure 4B) reflects heterogeneous expression across cells or low expression in each cell, we summed reads from all CSS genes within each individual cell. We found that there were less CSS transcripts overall in individual visual and olfactory KCs relative to non-KCs (Figure 5A). Using multiple methods, we failed to identify any further deterministic or non-deterministic variation in surface molecule expression among cells of the same class and stage.⁷² These analyses support the model that cell surface molecules are depleted in individual KCs.

Figure 5: — **(A)** Summed expression of 932 cell surface and secreted molecules in individual Kenyon cells and non-Kenyon cells sorted as false positives. Throughout this figure, violin plots show distributions of values for individual cells with medians and quartiles marked.

**(B,C)** UCell gene enrichment scores for Ig-SF’s (B) and GPCRs (C). N for each category shown in Figure 4G. Additional gene sets shown in Figure S5.

**(D)** tSNE of *fruitless*-expressing neurons at 48h APF from Palmateer et al. Labeled clusters are shown in (E,F). Additional analyses shown in Figure S6 to S8.

**(E,F)** UCell scores of Ig-SF and GPCR gene sets for select cell types and clusters from (D). Clusters are ordered by UCell median. Additional gene sets are shown in Figure S8.

See also Figures S5,S6, S7, S8.

We next examined expression of subgroups of CSS genes in KCs. The most depleted genes in the bulk RNAseq data (Figure 4B) were the Ig-SF set. Genes in this family are expressed in diverse and specific patterns across different types of neurons in the fly brain^{50,52,54-56,70,73-75} and are predicted to regulate synaptic partner choice or other type-specific processes during neuronal development.^53,54,76-83 We quantified the enrichment of Ig-SF molecules in each cell using UCell, a ranking-based approach for quantifying gene signature scores.⁸⁴ These specificity molecules were strikingly low-ranked in KCs of all types compared to non-KCs, and this relationship was true across time points (Figure 5B). KCs are small and have less RNA per cell than other central brain neurons, and we obtain fewer unique molecular identifiers (UMIs) per KC than for other cells (Figure S5A). UCell’s rank-based computation successfully accounts for this difference: Expression of transcription factors and G protein-coupled receptors (GPCRs) ranked similarly between KCs and non-KCs, as did expression of a different group of putative synaptic specificity molecules, leucine rich repeat proteins (LRRs) (Figure 5C, Figure S5 B,C). Interestingly, the entire family of beats was depleted (Figure S5D, S7B). 14 Beats form complex heterophilic receptor-ligand networks with eight Sides.⁷³

The non-KCs in our dataset represent a random and heterogeneous sample. We thus repeated our analysis in an scRNAseq dataset that covers fruitless-expressing neurons, including some KC types.⁸⁵ We re-analyzed this dataset using our standard methods and identified 132 clusters of neurons, including three types of KCs and 23 clusters of optic lobe neurons previously identified by transcriptomics (Figure 5D, see Methods for details and Figures S6, S7) ^70,86. All three types of KCs had markedly less Ig-SF molecules than all other fruitless clusters, including the 23 cell types from the optic lobe (Figure 5E-F, Figure S8). Again, no differences were observed for transcription factors, GPCRs, or LRRs (Figure S8). Together, these analyses of individual KCs support a model in which each KC is depleted of Ig-SF’s, rather than different cells robustly expressing different family members.

Of course, KCs do not lack partner specificity entirely: KCs of different types receive input from visual versus olfactory modalities^23,31,87, and indeed some Ig-SF molecules are differentially expressed in olfactory versus visual KCs (Figure S5D).⁷² Olfactory KC subtypes innervate distinct axonal lobes and connect with lobe-specific dopaminergic neurons and mushroom body output neurons. As described previously, DIP and Dpr Ig-SF proteins are required for these processes.^80,83 Consistent with this, we find that certain cell surface genes are selective for distinct olfactory KC subtypes (Figure S5D).⁷² As KCs of every olfactory subtype can receive inputs from every PN type (Figure 2D), we do not expect differences in transcription between olfactory KC types to explain the diversity of inputs from PNs.

When deprived of typical partners, Kenyon cells still synapse

Our connectomic and transcriptomic analyses suggest that KCs partner randomly with nearby PN boutons rather than using cell surface molecules to structure connectivity. We sought to test how KCs would respond to removal of a large swathe of their typical PN partners – could they connect promiscuously? To accomplish this, we killed half the PNs during development by driving expression of diphtheria toxin CBβ with VT033008-Gal4, which labels a defined PN subset including most of the anterodorsal lineage (Figure 6A,B).²¹ VT033008 PNs contribute an average of 234 boutons in control (Figure 6C-E). When VT033008 PNs are ablated, remaining PNs, labeled by VT033006-LexA driver and/or Choline Acetyltransferase (ChAT) immunostaining, increase their bouton production, such that ablated calyces lack only ~170 boutons on average (Figure 6E). We next looked at ~12 PNs from 3 glomeruli (DA1, DC3, and VA1d ) labeled by MZ19 QF (Figure 6F-I). In the ablation condition, an average of 2 MZ19 PNs were lost, but we observed a 60% increase in MZ19 boutons in the calyx. Thus in addition to loss of VT033008-labeled types, this condition also involved PN-type-specific rebalancing of calyx composition, such that MZ19 boutons go from contributing 6% of boutons in wild type to 14% of boutons in the ablation condition (Figure 6G,I).

Figure 6: — **(A)** Maximum intensity Z projections of antennal lobes in control animals and animals subjected to ablation of half of PNs.

**(B)** Cell counts for PNs labeled by VT033008, VT033006, or both in control or VT033008 ablation. Wilcoxon-rank sum test. Throughout this figure, grey bars depict medians. Number of hemispheres can be inferred from number of dots in each plot.

**(C)** Single slices of control and VT033008-ablation calyces.

**(D)** Distribution of boutons numbers of each type.

**(E)** Distribution of bouton types in each sample. Students’ T-test or Wilcoxon rank-sum.

**(F)** Maximum intensity Z projections of control and VT033008-ablated antennal lobes.

**(G)** Cell counts for PNs labeled by MZ19 in control and VT033008 ablation. Students’ T-test.

**(H)** Single slices of MZ19 PNs in control and VT033008 ablated calyces.

**(I)** MZ19 bouton counts in control and VT033008 ablation. Students’ T-test.

**(J-K)** Control and VT033008 ablation calyces labeled with an antibody against presynaptic marker Bruchpilot (Brp) (J) or postsynaptic marker Discs large (Dlg) (K).

**(L)** Area of the largest slice of the calyx with reference to ChAT (circular points) or Brp (triangular points) immunostaining from control and ablation brains. Student’s T-test.

**(M)** Normalized Brp intensity for control and ablation calyces. Wilcoxon rank sum test

**(N)** Normalized Dlg intensity for control and ablation calyces. Student’s T-test

KCs appear able to obtain a normal number of inputs in this condition, suggesting that bouton subtype identity does not restrict partnerships: We have previously reported that KC number is unchanged by the loss of VT033008 PNs, and that even extreme olfactory PN loss does not affect KC claw number.²¹ Here, we find that the size of the calyx and its synaptic density (measured by presynaptic Bruchpilot and postsynaptic Discs large) are robust to loss of these PNs (Figure 6 J-N). As individual boutons can have widely varying numbers of synapses and connect with widely varying numbers of KCs (Figure 2D), our data is consistent with a model in which spared PNs modestly increase their bouton number while also increasing the number of synapses on each bouton and connected claws to preserve the overall synaptic repertoire of the calyx.

Ectopic Ig-SF expression attracts Kenyon cell claws onto boutons of projection neurons expressing cognate molecules

We and others have shown that in the visual system, neurons experimentally depleted of their Ig-SF molecules become more promiscuous in forming partnerships.^53,54 We thus asked the opposite, if adding an Ig-SF protein to KCs could reduce their promiscuity. Among other Ig-SF’s, MZ19 PNs are enriched for expression of DIP-eta and DIP-theta throughout development (Figure 7A, B).⁵¹ KCs express little to none of the DIP-eta /theta partner, Dpr4 (Figure S5D).⁵⁵ We therefore drove ectopic expression of dpr4 in γ-KCs, which usually ramify broadly, and asked whether this could influence their overlap with MZ19 boutons versus neighboring, non-MZ19 boutons (Figure 7C-E). γ KC fluorescence increased on MZ19 boutons and was accompanied by a similar increase in the intensity of the postsynaptic marker Discs large (Figure 7F,G,H). These results suggest that ectopic Dpr4 expression in γ KCs biased them toward connecting with MZ19 partners, increased the number of synapses between γ KCs and their randomly chosen MZ19 partners, or both.

Figure 7: — **(A, B)** tSNE plots of developing PN scRNAseq from Xie et al. ⁵¹ In (A), clusters of cells belonging to PN types labeled by genetic driver MZ19 are colored according to their type and developmental stage. (B) shows expression of Dpr4 partners *DIP-eta*, *DIP-theta*, and *DIP-iota*.

**(C)** Single confocal slices of control calyx and calyx in which γ KCs ectopically express Dpr4. Dotted outline circles the calyx. Boxed region is magnified and shown in insets at right. Pink outlined bouton: MZ19-QF-positive. Cyan outlined bouton: MZ19-QF-negative.

**(D)** Average normalized fluorescence values for non-MZ19 (cyan) and MZ19 (pink) boutons of each calyx for the control and Dpr4 overexpression samples. Values from the same calyx are connected by a line. Wilcoxon signed rank exact test for values from the same brain and Students’ T-test for comparisons between brains.

**(E)** Cumulative distribution of γ KC fluorescence in individual boutons within each calyx. Each line plots the bouton fluorescence distribution within each calyx.

**(F)** Single confocal slices showing expression of synaptic proteins in control (left) and Dpr4OE (right) calyces. Representative MZ19 boutons circled in pink.

**(G)** Average normalized KC (89B01+) fluorescence values for MZ19 boutons in control and Dpr4OE calyces in brains also stained with an antibody against Dlg. These are different brains than those in (D). Student’s T-test.

**(H)** Average normalized Dlg fluorescence values for MZ19 boutons in control and Dpr4OE calyces. Student’s T-test.

**(I)** MZ19 bouton counts in control and Dpr4OE. Boutons were counted in two different experiments where MZ19QF boutons were recognized with antibodies against tomato (triangular points) or HA (circular points). Wilcoxon rank sum test.

**(J,K)** Visualization (J) and claw quantification (K) of single γ Kenyon cells electroporated with Texas red Dextran. Claw structures are marked by red arrows. Wilcoxon rank sum test.

To further interrogate these models, we counted the number of MZ19 boutons and used Texas red Dextran dye electroporation to measure the number of claws on individual γ KCs across conditions. We found that the median numbers of boutons per calyx and claws per cell were consistent across genotypes, while the variances increased (Figure 7I-K). The consistency in median claw and bouton numbers across conditions rules out a model in which dpr4 impacts absolute bouton stabilization, dendritic branching, or KC stickiness. Together, these results are consistent with our hypothesis that the natural lack of specificity molecules in KCs allows them to synapse promiscuously with molecularly diverse presynaptic PNs.

Discussion

In 1958, Frank Rosenblatt formulated a model learning circuit, the Perceptron, which bears an uncanny resemblance to the mushroom body. He proposed that the Kenyon-cell-like “A units” would receive random input combinations from the sensory apparatus.⁸⁸ Though modern neural networks have diverged from their biological inspiration in the brain’s actual intelligence, the utility of randomized initial conditions remains a key feature of the computational method. Of course, we now know that there is tremendous specificity in the distribution of neuronal cell types and connections in many parts of the brain, and while Rosenblatt assumed that inputs to neurons engaged in associative learning were randomized, he lacked the data to do so. Whether the inputs to biological learning circuits are in fact randomly wired has therefore been the subject of much debate.

While the odorant perception and olfactory associative learning functions of the mushroom body have been well established ^24,89-91, debate about the random or structured nature of the connectivity continues.^{6,18,22-25,31-40} In our developmentally-based model, we find that previously competing theories are both right—yes, cell types are structured in particular locations in the calyx, and this can give rise to regional correlations in KC inputs or odor representations; and yes, KCs appear to choose randomly from spatially available PNs.

This model of spatially constrained but nondeterministic input combinations could also explain patterns of inputs to other expansion layers. In the cerebellum, inputs vary across individual cells, but exhibit regional correlations.^5,7,44,45 Our model could also describe the regionalization of sensory inputs to the cortex and overlapping spatial domains in piriform even if cortical neurogenesis programs are themselves homogenous.^92,93 Moreover, just as KC clonal relationships predict odor input correlations, lineally related “sister” neurons structure cortical columns in and have correlated activity.^94,95 However, this mechanism may not describe all connections in all expansion layers: A connectomic analysis of neocortex from an adult mouse found that patterns of synapse formation could not be explained by random choice within spatial boundaries.⁹⁶

How do these patterns develop? We reason that KCs are selective for olfactory PNs as a group, but connect without molecular preference to nearby subtypes by making use of underlying promiscuity present in all neurons. In culture, neurons will synapse with whoever they touch, or with themselves, and the existence of “neuronal self-avoidance” systems (Dscams, clustered protocadherins) implies that neurons will form synapses unless told not to.^97-101 Other studies have unmasked promiscuous behavior in neurons with typically stereotyped partners by altering filopodial dynamics or displacing neurites.^102-104 The cell adhesion molecules lacking in KCs constrain partner choice elsewhere, but are not required for synaptogenesis per se: When they are removed from neurons that naturally express them, connections become more promiscuous.^53,54,105

While the Ig-SF depletion of KCs is correlated with promiscuity, other mechanisms may also contribute. First, there may be sources of cell autonomous molecular variation among KCs that we have not yet discerned. Second, it is possible that PN surfaces are more similar than their transcriptional profiles would suggest, due either to translational control or trafficking. Finally, the impact of Ig-SF depletion could be causal, but act primarily on the biophysical characteristics of the growth and stabilization of KCs; this could allow for unique modes of input acquisition, irrespective of molecular matching.

We are not suggesting that KC development or connectivity is willy-nilly. First, while olfactory and visual KCs both connect promiscuously, they each do so only within precise groups of partners.^23,31 This could arise from molecular determinants common to their presynaptic partners as groups or from spatial or temporal stratification of these neuronal pools. Second, though KCs as a population can connect to all kinds of olfactory PNs, individual KCs receive only a handful of inputs, and the choices of individual cells are constrained by their reach. Limited KC dendrite spread could be a cell autonomous feature or could result from their rapidly meeting a viable partner. As KCs are most restricted along the M-L axis, the distribution of PN boutons along this axis determines the local mix of potential inputs.³⁹ PN bouton M-L position seems to be linked to the specific number of collaterals made by the cell, though how this occurs is unclear.

Our model has the benefit of allowing both developmental robustness and evolvability: If postsynaptic cells are not choosey, the circuit can flexibly incorporate additions or changes in projection neuron inputs, as we demonstrated with our ablation experiments here and previously.²¹ Over evolutionary time or across different expansion layers, changes in the organization of inputs could produce a wide variety of circuit functionalities: Well-mixed inputs could produce a mathematically ideal Marr-Albus expansion layer, while inputs mapped along a tonotopic, spatial, or chemical axis would be reflected in structured correlations in the tuning of expansion layer neurons.¹⁰⁶

Given the large number of neurons in expansion layers, the compactness of the developmental information required to wire them is a critical aspect of fitness. For example, the mammalian cerebellum can comprise tens of billions of granule cells; a finite genome cannot deterministically specify unique inputs to each of these billions of cells. Because meaning need not be hard coded in learning circuits, non-deterministic connectivity is compatible with function.^88,107-109 The “subtype-agnostic” system through which we hypothesize Kenyon cells acquire their olfactory presynaptic partners provides instructional economy. The simple principles we describe are easily scalable and could explain the development of vastly larger expansion layers in other organisms.

STAR Methods

Experimental Model and Study Participant Details

Fly husbandry

All fly lines used in this study are listed in the key resources table. Fly genotypes by figure can be found in Table S2. Flies were maintained on Bloomington food with a yeast sprinkle (“B” recipe, Lab Express, Ann Arbor, MI) at 25C on a 12:12 light-dark cycle with at least 60% humidity (provided by a beaker of water in the incubator). Flies for bulk RNAseq were maintained on Rockefeller University molasses food. For staging developmental time points, pupae were marked at the 0h APF “white pupae” stage. Animals for scRNAseq were collected +/−2 hours from the stated developmental time points (e.g. 48h APF were collected from 46-50h APF). Pupae for bulk RNAseq were collected at 44-46h APF. Flies for adult scRNAseq were maintained on Desplan lab standard cornmeal food at 25C on a 12:12 light-dark cycle without controlling for humidity. Females for adult scRNAseq were collected <2 weeks after eclosion.

Key Resources Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
anti-ChAT (mouse monoclonal)1:200	Developmental Studies Hybridoma Bank	Cat# chat4b1; RRID: AB_528122
anti-DsRed (rabbit polyclonal) 1:1000	Clontech/Takara Bio	Cat# 632496; RRID: AB_10013483
anti-GFP (chicken polyclonal) 1:5000	Cai Lab	N/A
anti-brp (mouse monoclonal) 1:40	Developmental Studies Hybridoma Bank	Cat# nc82; RRID: AB_2314866
anti-GFP (rabbit polyclonal) 1:250	Fisher	Cat# A11122; RRID: AB_221569
anti-RFP (rat monoclonal) 1:250	Chromotek	Cat# 5f8; RRID: AB_2336065
anti-HA (Rat) 1:100	Millapore Sigma	Cat# 11867423001; RRID: AB_390918
anti-dlg (mouse monoclonal) 1:500	Developmental Studies Hybridoma Bank	Cat# 4f3; RRID: AB_528203
Alexa 488 anti-chicken (goat polyclonal) 1:1000	Fisher	Cat# A32931; RRID: AB_2762843
Alexa 488 anti-rabbit (goat polyclonal) 1:1000	Fisher	Cat# A11008; RRID: AB_143165
Alexa 568 anti-rabbit (goat polyclonal) 1:1000	Fisher	Cat# A-11036; RRID: AB_10563566
Alexa 568 anti-rat (goat polyclonal) 1:1000	Fisher	Cat# A11077; RRID: AB_2534121
Alexa 647 anti-mouse (goat polyclonal) 1:1000	Fisher	Cat# A21236; RRID: AB_2535805
Chemicals, peptides, and recombinant proteins
Paraformaldehyde	EMS	15710
DAPI	Sigma	D9542
10X PBS Gibco	Fisher	70-011-044
Normal Goat Serum	MP biomedicals	8642921
Triton	Sigma	X100
Sodium chloride NaCl	Sigma	S7653
Potassium chloride KCl	Fisher	P217-500
Calcium chloride CaCl2*2H2O	Fisher	C70-500
Magnesium chloride MgCl2*6H2O	Sigma	M2670
HEPES sodium salt	Sigma	H7006
Trehalose	Sigma	T9531
Sucrose	Sigma	84097
Sodium bicarbonate NaHCO3	Fisher	S233-500
Sodium phosphate NaH2PO4*H2O	Fisher	S369-500
Propyl gallate	Sigma	02370
Xylenes	Fisher Chemical	Cat# X5-1
DPX	Sigma Aldrich	Cat# 06522-100ML
Texas red Dextran	Thermo Fisher	Cat# D3328
Schneider's Drosophila Medium	Gibco	21720024
Collagenase from Clostridium histolyticum	Sigma	C0130
Dispase II	Sigma	D4693
Actinomycin D	Sigma	A1410
1X Dulbecco’s PBS	Corning	21-031-CV
50 mg/mL UltraPure BSA	Invitrogen	AM2618
5 mM DRAQ5 Fluorescent Probe Solution	Thermo Scientific	62254
Deposited data
Bulk RNA sequencing of Kenyon cells and projection neurons	This study	GEO: GSE276028
scRNAseq of Kenyon cells at developmental time points	This study	GEO:GSE296723
scRNAseq of adult Kenyon cells	This study	GEO: GSE274592
Experimental models: Organisms/strains
GH146 Gal4 (II)	Stocker et al., 1997¹¹⁶	BDSC 30026
VT033008-Gal4 (attp2)	Yoshi Aso, Janelia Farms Research Campus¹¹⁷
89B01-Gal4 (attp2)	Jennett et al., 2012¹¹⁸	BDSC 40541
OK107Gal4 (IV)	Connolly et al., 1996¹¹⁹	BDSC 854
10XUAS-IVS-myr::tdTomato (attp40)	Pfeiffer et al., 2012¹²⁰	BDSC 32222
UAS-CD8GFP (III)	T. Lee et al., 1999¹²¹	BDSC 5130
10XUAS-IVS-myr::GFP (attP2)	Pfeiffer et al., 2012¹²⁰	BDSC 32197
UAS-Cbβ\DT-A.I[18] (II)	Han et al., 2000¹²²	BDSC 25039
UAS luc RNAi (III)	Zirin et al., 2020¹²³	BDSC 35788
UAS DPR4 OE (86Fb)	Bischof et al., 2013¹²⁴	FlyORF F002762
VT033006:: LexAp65 (JK22C)	Heather Dionne and Gerry Rubin, Janelia Farms Research Campus^125,126
91G04-LexA, LexAopTom (attp40)	Liqun Luo, Stanford University¹²⁷
LexAopTomato (suHw attp5)		FBti0160868
MB247-LexAVP16 (III)	Scott Waddell, University of Oxford¹²⁸
LexAop CD2-GFP (III)	Lai & Lee, 2006¹²⁹	BDSC 66544
GH146QF, QUAS-mtdTomato-3xHA (II)	Potter et al., 2010¹³⁰	BDSC 30037
MZ19-QF (II)	Hong et al., 2012¹³¹	BDSC 41573
QUAS-mtdTomato-3xHA (II)	Potter et al., 2010¹³⁰	BDSC 30004
P(caryP) attp40	Markstein et al., 2008¹³²	BDSC 36304
Empty landing site in 86FB (removed 3XP3-RFP with CRE)	Bischof et al., 2013¹²⁴	Fly ORF/FBti0076525
OK107-QF2>QUAS-GFP, UAS-mtdTomato		BDSC 66472
P{y[+t7.7] w[+mC]=R13F02-p65.AD}attP40/CyO; P{y[+t7.7] w[+mC]=R89B01-GAL4.DBD}attP2	Aso et al., 2014¹¹¹	BDSC 68265
P{y[+t7.7] w[+mC]=R26E07-p65.AD}attP40/CyO; P{y[+t7.7] w[+mC]=R30F02-GAL4.DBD}attP2	Aso et al., 2014¹¹¹	BDSC 68322
P{y[+t7.7] w[+mC]=R35B12-p65.AD}attP40; P{y[+t7.7] w[+mC]=R34A03-GAL4.DBD}attP2/TM6B, Tb[1]	Aso et al., 2014¹¹¹	BDSC 68370
P {y[+t7.7] w[+mC]=R13F02-p65.AD}attP40; P{y[+t7.7] w[+mC]=R85D07-GAL4.DBD}attP2	Aso et al., 2014¹¹¹	BDSC 68383
P{y[+t7.7] w[+mC]=R13F02-p65.AD}attP40; P{y[+t7.7] w[+mC]=R58F02-GAL4.DBD}attP2	Aso et al., 2014¹¹¹	BDSC 68255
P{y[+t7.7] w[+mC]=R52H09-p65.AD}attP40; P{y[+t7.7] w[+mC]=R18F09-GAL4.DBD}attP2	Aso et al., 2014¹¹¹	BDSC 68267
GAL4^{dpr1-MI02201-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349605
GAL4^{dpr3-MI05963-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349598
GAL4^{dpr5-MI11085-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349594
GAL4^{dpr6-MI01358-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349602
GAL4^{dpr8-MI06778-TG4.0}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349599
GAL4^{dpr9-MI03594-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349603
GAL4^{dpr10-MI03557-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349595
GAL4^{dpr11-MI01743-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349596
GAL4^{dpr12-MI01695-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0349600
GAL4^{dpr15-MI01408-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0328010
GAL4^{DIP-α-MI02031-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0345530
GAL4^{DIP-δ-MI08287-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0345531
GAL4^{DIP-η-MI07948-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0345533
DIP-γ^MI03222-GAL4	Zinn and Zipursky labs, via Pelin Volkan	FBal0319064
DIP-κ^MI01295-Gal4	Zinn and Zipursky labs, via Pelin Volkan	No record, generated from FBal0264429 by Carillo et al
GAL4^{DIP-θ-MI03191-TG4.1}	Zinn and Zipursky labs, via Pelin Volkan	FBal0345534
GAL4^{DIP-ζ-MI03838-TG4.0}	Zinn and Zipursky labs, via Pelin Volkan	FBal0345532
Software and algorithms
FIJI	Schindelin et al., 2012¹³³	https://imagej.net/software/fiji/downloads
R	R Foundation for Statistical Computing, 2023¹³⁴	https://www.R-project.org/
Neuprint: An open access tool for EM connectomics	Clements et al., 2022¹³⁵	https://neuprint.janelia.org/
R package natverse	Bates et al., 2020¹³⁶	https://natverse.org/
R package Seurat	Hao et al., 2021; Satija et al., 2015; Stuart et al., 2019^137-139	https://satijalab.org/seurat/
R package Clustree	Zappia & Oshlack, 2018 ¹¹²	https://lazappi.github.io/clustree/
R package UCell	Andreatta & Carmona, 2021⁸⁴	https://github.com/carmonalab/UCell
R package RColorBrewer	Neuwirth, 2022 ¹⁴⁰	https://CRAN.R-project.org/package=RColorBrewer
python	Python Software Foundation	http://www.python.org
python package scanpy	Wolf et al., 2018¹⁴¹	https://github.com/scverse/scanpy
python package scvi-tools	Gayoso et al., 2022¹⁴²	https://github.com/scverse/scvi-tools
Other
Fly food	Lab Express, Ann Arbor MI	“B” Food
Cell strainer caps (used as colanders for stainings)	Fisher	BD 352235
Binder reinforcement stickers (for mounting brains)	Office Depot	Avery 5722
pluriStrainer Mini 20 μm	pluriSelect	43-10020-40
10cm Borosilicate glass capillary tubes with filament	Sutter Instruments	Cat #BF100-50-10
Stimulator	Grass	model: SD9
Tungsten Pins	California Fine Wire Company	Cat# 100211

Open in a new tab

Method Details

EM analysis

Downloading synapse locations from neuprint+

Cartesian coordinate locations of synapses between uniglomerular projection neurons and Kenyon cells from the Hemibrain: v1.2.1 were downloaded from neuprint+ using a custom query which retrieved location of synapses between 115 uniglomerular projection neurons and any cell with ‘KC’ in its instance (name). The unique Hemibrain cell body Id for both pre and postsynaptic partners was also included in the output. The custom query was run 6x with batches of about ~20 PN cell body Ids at a time.

Example query:

MATCH(pn:Neuron)-[:Contains]->(:`SynapseSet`) -[:Contains]->(pns:Synapse)-
[:SynapsesTo]->(kcs:Synapse)<-[:Contains]-(:`SynapseSet`)<-[:Contains]-(kc:Neuron)
WHERE pn.bodyId IN [1639234609, 818983130, 1796818119, 1734350788, 1788300760,
5813054697, 5813024710, 5813024601, 754534424, 1670934213, 1734350908, 5813039315,
722817260, 1888572074, 5813055184, 1609542060, 5813039465, 850375847, 1914140664,
5813024712] AND kc.instance=~'KC.*' AND kc.status = "Traced" AND pns.type = "pre"
AND pns.`CA(R)`
RETURN pn.bodyId, pn.instance, kc.bodyId, kc.instance, pns.location.x, pns.location.y,
pns.location.z, pns.type

The list of 115 cell body Ids for uniglomerular projection neurons was obtained by using the “find neurons” function in neuprint+ to search for neurons with inputs in AL(R) and outputs in CA(R). The output of 136 neurons with ‘PN’ in their instance was further filtered to exclude 21 multiglomerular PNs using a list of Hemibrain counterparts of PNs reconstructed from the FAFB by Bates et al. In the Hemibrain, only 112 of these PNs made synapses with KCs in the calyx. For all analyses the 112 PNs were filtered to 105 olfactory PNs by removing 7 thermo- and hygro-sensory PNs innervating the ventral posterior glomeruli- VP1d, VP1m, VP2, VP3, VP4, and VP5.

25,718 unique presynapses from uniglomerular olfactory projection neurons to KCs were found. Monosynaptic connections between PNs and KCs were disproportionally present in the data and so filtered out as false positives down to a total of 25,394 presynapses. For all analyses only synapses between olfactory PNs and main calyx KCs were considered. These amounted to a total of 24,791 synapses.

Downloading, and trimming neuron skeletons

All PN and KC skeletons were downloaded using the “neuprint_read_neurons” function in the neuprintR R package from our list of body Ids from PNs and KCs. Skeletons were roughly trimmed to a cube surrounding the calyx which was 5% larger on each end of the x and z axis and 10% larger on the y axis than the bounds of the CA(R) mesh (neuprint_ROI_mesh(‘CA(R)’)).

Cell type and developmental feature annotation

Each PN was annotated with an AL glomerular target/ cell type and neuroblast origin in neuprint+ and in Bates 2020 ^46,57. We further added developmental stage of birth and birth order information from Yu et al 2010 and Lin et al 2012 ^29,30.

We used KC type annotations from neuprint+ ⁴⁶. γ KCs were KCs with instances including ‘KCg-m’,’KCg-d’, ‘KCg-t’, ‘KCg-s’, and ‘KCy(half). α’β’ KCs were KCs with instances including 'KCa’b’-m', 'KCa’b’-ap1', and 'KCa’b’-ap2'. αβ KCs were those with instances including 'KCab-s','KCab-m','KCab-c','KCab-p'. Unknown KCs included instances of ‘unknown’, ‘KC part due to gap’, ‘KC (incomplete?)’. Kenyon cells were filtered to only the 1728/ 1874 that innervate the olfactory calyx (not γ-d, α’β’ ap1, nor αβp).

We assigned each KC to one of 4 neuroblasts of origin by using the location of its axon in the pedunculus where axons are known to segregate in a clonal fashion ²⁸. We started by assigning putative neuroblast origin to the pupally born αβ KCs. As the latest born KC subtype, their axons are surrounded by α’β’ and γ KCs born from their same neuroblast and so are most separated from αβ cells born from different neuroblasts. We used k-means clustering to sort points of intersection between each KC and each of three planes perpendicular to the pedunculus into 4 groups. 806/809 αβ KCs were assigned the same cluster at each of the three planes and so were sorted into neuroblasts named A,B,C,D from medial to lateral.

We next assigned α’β’ KCs to neuroblasts by again calculating the points of intersection between axons and planes perpendicular to the pedunculus at 3 different levels. Rather than using k-means clustering of points of intersection; each unassigned KC was assigned to the clone to which it had the lowest mean distance from each of the axons already assigned. 246/336 α’β’ KCs were automatically sorted into neuroblasts in this manner. Finally, the earliest born γ KCs were sorted using the same method proximity based method to sorted αβ and α’β’ KCs. 264/593 γ KCs were automatically assigned using this method. KCs of all three types which were not able to be computationally assigned to clones using axon intersections in the pedunculus were manually assigned to clones based on their soma position which is also known to be sorted by clones ²⁸ and further proofread with pedunculus position. The number of cells of each type assigned to each clone is roughly equivalent, as expected (Figure 2A).

Subcellular feature annotation

Using the locations of synaptic contacts, we obtained spatial information about sub-cellular features of both projection neurons and Kenyon cells in the calyx. For projection neurons, we estimated coordinate centers of boutons and of collateral branch points from the main axon. For Kenyon cells, we estimated coordinate centers of claws.

Bouton centers were calculated as the coordinate center of the presynapses that made up the bouton. Presynapses clustered together most often at the ends of collaterals into putative boutons. We used k-means clustering to quickly assign groups of presynapses to individual boutons using the number of putative boutons as the number of clusters. We next manually ‘proof-read’ bouton assignments reassigning clear errors, e.g. presynapses clearly on other collateral branches or not immediately next to other presynapses assigned to that bouton. In sum we found 538 boutons which matches with our counts in Figure 6D (median = 524.5 boutons). We limited analyses to 436 olfactory PN boutons that were made up more than 3 presynapses.

Boutons most often are at the ends of collaterals which branch perpendicularly from the PN axon in the medial Antennal Lobe Tract (mALT). We extracted the locations of these collateral branch points by filtering branch points in the PN skeleton to those on the segment defined by the natverse function ‘spine’ (the longest continuous track) which was the axon in the mALT and then manually assigned them by plotting putative branch points onto the skeleton with bouton centers. Collateral branch points were branch points at the base of collaterals that bore boutons. In sum we annotated 273 collateral branch points.

Each Kenyon cell claw center was calculated as the coordinate center of the presynapses of a single PN bouton opposed by a single Kenyon cell. In sum we annotated 10,970 Kenyon cell claws. We limited our analyses to 10,674 claws belonging to main calyx Kenyon cells. Kenyon cell claws that opposed fewer than 3 presynapses were also filtered out reducing their number to 10,195 in the olfactory calyx.

Rotating coordinate space to match up with relevant calyx axes

To simplify analyses of the calyx we rotated all coordinates such that the longest extent of the calyx along which the mALT runs (M-L) was parallel to the X axis and the shortest axis (A-P) perpendicular to the mALT was parallel to the Y axis. Coordinates were rotated 285 degrees about the X axis and 200 degrees about the Y axis.

Generation of random models of KC:PN connectivity

In order to ask if patterns of non-random connectivity between PNs and KCs could be explained by developmentally defined spatial limits on the neurites of both PNs and KCs, we generated a series of null models of connectivity to compare with observed connectivity in the Hemibrain. We considered 20 different developmentally defined groups of Kenyon cells: all olfactory Kenyon cells, Kenyon cells belonging to each of the 4 clones, Kenyon cells of each of the 3 types, and Kenyon cells belonging to each of the 12 groups of clone and type. For each group we generated 10,000 null models of connectivity in which each claw of each Kenyon cell is randomly assigned to a PN type. The probability of selecting any PN type was proportional to the number of boutons accessible to KCs of that group in the Hemibrain dataset.

Conditional Input analysis

We performed conditional input analysis as described in Zheng 2022 ²⁵ on connectivity matrixes for each of our 20 developmentally defined KC groups. The output of this analysis is a matrix of PN types (51x51), which indicates the number of KCs connected to a column PN type given they also are connected to a row PN type. By comparing values computed using observed Hemibrain connectivity to distributions of values computed using the 10,000 random models we can assess how similar observed connectivity is to that of each of the random models.

Bulk RNA sequencing

Flow cytometry

Flow cytometry was performed largely as described in ¹¹⁰. Brains were dissected for up to 60 minutes in Schneider’s medium supplemented with 1% BSA and placed on ice. Optic lobes were removed during dissection. After all dissections were completed, collagenase was added to a final concentration of 2mg/mL and samples were incubated at 37C for 20 minutes (adults) or 12 minutes (pupae), without agitation. Samples were dissociated by trituration and spun down at 300g, 4C, for 5 minutes, in a swing-out rotor. Collagenase solution was removed and replaced with PBS+0.1% BSA, and cells were passed through a cell strainer cap and supplemented with 50ng/mL DAPI before being subjected to flow cytometry on a FACS Aria II (University of Michigan Flow Cytometry Core for scRNAseq, Rockefeller University Flow Cytometry Core for bulk RNAseq). Plasticware for cell dissociation and collection was pre-treated by rinsing with PBS+1% BSA to prevent cells from sticking to bare plastic.

During flow cytometry, dead and dying cells were excluded using DAPI signal, and forward scatter and side scatter measurements were used to gate single cells. Using our dissociation methods, 50–90% of singlets appeared viable (DAPI-low). Fluorophore-positive cell rate varied with the prevalence of the cell population. During sorting, we made two adjustments to protect the fly primary cells, which were very delicate—we disabled agitation of the sample tube, and sorted using the “large nozzle,” e.g. 100μm, i.e. using larger droplet size and lower pressure. For bulk RNAseq, we sorted cells directly into Trizol-LS. For scRNAseq, we sorted cells into PBS+1% BSA.

For adult scRNA-seq, brains were dissected for up to 60 minutes in Schneider’s medium and placed on ice. Optic lobes were removed during dissection. After all dissections were completed, collagenase and dispase were added to a final concentration of 2mg/mL each along with 45 uM Actinomycin D, and samples were incubated at 26C for 1 h without agitation. Samples were washed 2-3 times with DPBS+0.04% BSA, dissociated by trituration and passed through a 20 um PluriStrainer. Cells were stained with 5 uM DRAQ5 before being subjected to flow cytometry on a FACS Aria II (New York University Genomics Core). Forward scatter area and width were used to gate single cells, and cells were sorted into DPBS+0.04% BSA.

RNA isolation, library preparation, and sequencing for low-input bulk RNAseq on FAC-sorted neurons

To collect olfactory projection neurons and Kenyon cells for bulk RNAseq, we dissected, removed optic lobes, dissociated, and FAC-sorted about 100 brains (adult) or 25 brains (45h APF) for each replicate in which Kenyon cells were labeled by MB247LexA>LexAopGFP, and olfactory projection neurons by GH146-Gal4>UAS-mTdTomato. GFP-positive cells were 5-8% of viable singlets, and Tomato-positive were 0.4-1% of viable singlets. These numbers are consistent with the rate at which these cell types appear in the central brain (MB247+ Kenyon cells: ~4000 out of ~50,000; GH146+ PNs: 200-300 cells out of 50,000).

RNA isolation was performed as we described in detail previously ¹¹⁰. Briefly, we FAC-sorted cells directly into Trizol-LS and stored at −80C prior to RNA preparation. Cells obtained in each replicate are shown in the table below. We followed the standard Trizol-LS protocol until the aqueous phase was isolated. We then passed the aqueous phase over Arcturus Picopure columns, including a DNAse treatment on the column, using modifications of the standard Picopure protocol described in ¹¹⁰. We quantified total RNA using a BioAnalyzer and aimed for 1ng starting material per library. As insect processing of rRNA makes the “RNA Integrity Number” irrelevant, we visually inspected total RNA profile to assess RNA quality.

For each replicate, we matched the amount of total RNA used for library prep across the three cell types, i.e. we matched Kenyon cell and “double negative” RNA amounts to the amount we obtained for olfactory PNs. Total rRNA was poly-A selected and libraries were prepared using SmartSeq V2 at the Rockefeller University Genomics Core and sequenced to a depth of 25 million reads on Illumina platforms as described in the Table S3.

Single-cell RNA-sequencing

Flow cytometry for single-cell RNA-sequencing of developing Kenyon cells

To compensate for the effects of different strain backgrounds and driver lines, we labeled Kenyon cells with MB247-LexA (a transgene containing the Mef2 promoter) or with OK107-Gal4 (a Gal4 insertion into the eyeless locus). We then FAC-sorted and performed scRNAseq at 36h, 48h, and 60h APF (one OK107 and one MB247 sample at each time point). Preparation for FACS was as described above in the flow cytometry methods section. Each time point was dissected on a separate day. We collected 40-80 brains per dataset.

To make sure all Kenyon cells were collected, regardless of fluorescence intensity, we used a loose gating protocol in FACS prior to scRNAseq. This was especially important for the MB247 driver, which labels different Kenyon cells with different intensity. The sorted populations therefore also contained small proportions of “non-Kenyon cells” that were sorted as false positives; we made use of these cells to compare expression of various groups of genes to that of Kenyon cells. Prior to library prep for sequencing, cells were centrifuged at 4°C to increase the cell concentration per microliter.

Labeling of adult Kenyon cells and preparation for scRNAseq

To generate an adult Kenyon cell data set and annotate the seven molecularly defined cell subtypes ¹¹¹, we single-cell sequenced seven genotypes that simultaneously expressed GFP in all Kenyon cells and mtdTomato in a defined Kenyon cell subtype(s). To generate such flies, the OK107-QF2>QUAS-GFP, UAS-mtdTomato line (BDSC# 66472) was crossed with each of the following subtype-specific split-GAL4 lines ^66,111: γm and γd (BDSC #68265), γd (BDSC #68256), α′/β′m (BDSC #68322), α′/β′ap (BDSC #68370), α/βp (BDSC #68383), α/βc (BDSC #68255), α/βs (BDSC #68267).

The brains of adult female F1 progeny with the respective OK107-QF2>QUAS-GFP, GAL4-AD∩GAL4-DBD>UAS-mtdTomato genotype were dissected, visually checked for mtdTomato fluorescence in the mushroom body, dissociated into single cells, and stained with the DNA dye DRAQ5. DRAQ5+,GFP+ cells were FAC-sorted to enrich the samples for Kenyon cells. mtdTomato expression was not used for sorting. Given the prior knowledge that different classes of Kenyon cells separate well during the analysis of single-cell data ⁶⁴, the brains of flies where the subtypes of different classes were labeled (e.g. γd, α′/β′ap, and α/βs) were pooled in different combinations before the dissociation to minimize batch effects. Each brain pool corresponds to one library and includes at least 8 brains. In total, 8 single-cell libraries were prepared using the 10X Chromium Next GEM Single Cell 3’ Kit v3.1. The libraries were sequenced on either Illumina NextSeq500 or NovaSeq6000 to the depth of at least 28k reads/cell. All the analyses were performed using scanpy and scvi-tools. After the initial round of clustering, clusters containing Kenyon cells were selected based on the expression of dac, ey, and Pka-C1. Then, Kenyon cells were re-clustered and class identities were assigned based on the expression of the γ marker ab, the α′/β′ marker CG8641, and the α/β marker Ca-alpha1T. Next, subtype identities were assigned using the following criteria: 1) expression of mtdTomato transcripts (for example, if library 1 contained brains expressing mtdTomato under a γd-specific split-GAL4 driver, the γ cluster expressing mtdTomato in this library was designated as γd), 2) gene expression correlation between pseudo-bulk data generated from each cluster and bulk data from sorted cells expressing fluorescent marker driven by subtype-specific split-GAL4s ⁶⁶, and 3) relative size of each cluster as compared to the known relative size of each subtype population ¹¹¹.

ScRNAseq prep, quality control metrics and clustering analysis of developing Kenyon cells

ScRNAseq data was generated using gel bead microfluidics (10x Chromium 3’ system) on a NovaSeq 6000 instrument, from FAC-sorted cells. Cell viability and aggregation was checked and number of cells to load was selected to ensure a low multiplet rate (~3%) and ample expected cell recovery (at least 3000 cells). Raw fastq read files were processed using Cell Ranger (v6.1.0 or v7.1.0) with default parameters. A custom reference was built from the Drosophila FlyBase reference genome and transgene sequences (Gal4, LexA, QF, GFP, mtdTomato). For each sample, the sequencing was effective in recovering between ~3,000-14,000 cells, with ~60-80% sequencing saturation, ~40,000-100,000 mean reads per cell and ~1600-2800 median genes per cell. UMI and cell barcode filtering was done to remove duplicate reads and non-cell-associated barcodes. A summary of metrics of acquired datasets can be found in Table S4.

In the resulting gene-barcode matrix, each element gives UMI count associated with a gene (row) and a barcode (column). All steps of single-cell RNA-seq analysis were performed using Seurat (v4.3.0 and above). Cells were included if they had at least 1,000 but no more than 5,000 genes detected, and less than 5% mitochondrial gene content. Cells from all time points and driver lines (OK107 and MB247) were merged together with data from a-priori-labeled adult Kenyon cells into one Seurat object. Gene counts per cell were normalized to total read counts per cell (NormalizeData function, LogNormalize method). Variable genes were identified for each dataset independently (FindVariableFeatures function), followed by selection of 2000 genes that were repeatedly variable across datasets (SelectIntegrationFeatures function). To perform integration (IntegrateData function), this gene set was first filtered to exclude mitochondrial genes, heat shock proteins, transgenes (e.g. Gal4, GFP etc. – this was done to prevent transgenic variation driving the clustering), and ribosomal protein genes. Integrated data was then scaled and PCA was run with 50 principal components (ScaleData and RunPCA functions). The ElbowPlot tool was used to assess the level of transcriptional variation explained by each principal component; based on this analysis, 40 dimensions were then selected to cluster cells (RunUMAP, FindNeighbors, FindClusters functions). In FindClusters, clustering resolution was selected as 0.4 after testing a range of resolutions and using the R package “clustree” to assess stability scores of clusters at different resolutions ¹¹². Neurons, glia, and Kenyon cells were identified based on expression of known markers in the literature ^64-66. As adult Kenyon cell classes were defined by a priori Gal4 classification of cells (described above), we could observe that the integrated data was coherent with the adult classification of clusters. In addition, libraries from the two drivers, MB247 and OK107, mixed well among each other and the transgenic backgrounds were not the key factor influencing the clustering (Figure 4C).

Selection of gene sets

Cell Surface and Secreted Molecules (CSS’s)

This is a published list of 932 cell surface and secreted molecules (CSSs) with cell recognition properties ⁶².

Ig-SF and LRR specificity molecules

To define a category of cell adhesion molecules from the much bigger list of CSS’s, we used gene classifications based on extracellular domain information for protein-coding genes from the Drosophila melanogaster extracellular domain database ¹¹³. This database has been used in work describing the transcriptional programs of synaptic specificity such as in the fly visual system ⁷⁰. Cell adhesion molecules were defined from protein domains of Ig (132 genes) and LRR (72 genes).

GPCRs

This is a list of 112 genes from Flybase; 111 of these were expressed in the Kenyon cell sequencing data, and 108 in the Palmateer data. 5 GPCRs that were also LRRs were excluded from this list: CG44153, Lgr1, Lgr3, Lgr4, rk.

Transcription factors

This is a list of 628 transcription factors from Flybase. 613 of these were expressed in the Kenyon cell sequencing data, while 600 were expressed in the Palmateer data.

MiMIC reporter validation of sequencing data

Expression pattern of CSS’s in the bulk- and single-cell sequencing datasets was cross-checked with expression patterns of MiMIC reporter lines for an immunoglobulin superfamily of proteins called dprs (defective proboscis response) and their binding partners, DIPs (dpr-interacting proteins). This group of genes was selected as they have known roles in synaptic matching and MiMIC lines were available for 14 of these genes ^52,53,74,80. As a readout for reporter expression, we used the mushroom body pedunculus, a region where the Kenyon cell axons converge forming a stereotyped pattern with easily identifiable KC types, with early-born γ axons at the margin and latest-born αβ axons in the center of the pedunculus (transverse section) ¹¹⁴.

Brain dissection, immunostaining, and confocal imaging

For immunostaining, brains were dissected in external saline (108 mM NaCl, 5 mM KCl, 2 mM CaCl2, 8.2 mM MgCl2, 4 mM NaHCO3, 1 mM NaH2PO4, 5 mM trehalose, 10 mM sucrose, 5 mM HEPES pH7.5, osmolarity adjusted to 265 mOsm) for up to thirty minutes before being transferred to paraformaldehyde in PBS. All steps were performed in cell strainer baskets (caps of FACS tubes) in 24 well plates, with the brains in the baskets lifted from well to well to change solutions or with brains mounted to Poly-L-Lysine coated coverslips in single wells of 24 well plates. Brains were fixed overnight at 4C in 1% PFA in PBS or for 25 minutes at room temperature in 4% PFA. In general, the 4% PFA method preserves neuronal structures better, while the 1% PFA method allows better antibody penetration. Some antibodies also work better with the 1% method. After same-day (4%) or overnight (1%) fixation, brains were washed 3x10’ in PBS supplemented with 0.1% triton-x-100 on a shaker at room temperature, blocked 1 hour in PBS, 0.1% triton, 4% Normal Goat Serum, and then incubated for at least two overnights in primary antibody solution, diluted in PBS, 0.1% triton, 4% Normal Goat Serum. Primary antibody was washed 3x10’ in PBS supplemented with 0.1% triton-x-100 on a shaker at room temperature, then brains were incubated in secondary antibodies for at least two overnights, diluted in PBS, 0.1% triton, 4% Normal Goat Serum. DAPI (1 microgram/mL) was included in secondary antibody mixes. Antibodies and concentrations can be found in the resources table.

Brains were mounted in 1x PBS, 90% glycerol supplemented with propyl gallate in binder reinforcement stickers sandwiched between two coverslips. Samples were stored at 4C in the dark prior to imaging. The coverslip sandwiches were taped to slides, allowing us to perform confocal imaging on one side of the brain and then flip over the sandwich to allow a clear view of the other side of the brain. Brains fixed on Poly-L-Lysine coverslips were mounted with DPX. Before mounting samples were post-fixed for 4 hours in 4% PFA, washed 7x for 10minutes in increasing concentrations of ethanol (30%,50%,75%, 95%, 2x 100%), and 3x for 5 minutes in xylenes. Samples dried for a minimum of 2 days before imaging and were stored at room temperature. Scanning confocal stacks were collected along the anterior-posterior axis on a Leica SP8 with 1 micrometer spacing in Z and ~150nm axial pixel size, using a 40x 1.3 NA objective.

Dye filling Kenyon cells

Brains from <1 day old flies were dissected, treated with collagenase for 30 seconds, and pinned to a sylgard plate. Pulled glass electrodes were filled with dye and connected to a stimulator. Under a 2-photon microscope electrodes were then positioned in the soma of 89B01-Gal4>UAS-GFP expressing Kenyon cells. A single voltage pulse (50V for 0.5ms) was applied. Filled Kenyon cells were imaged 10 minutes later to allow for the distribution of dye among the processes.

Quantification and Statistical Analysis

Analysis of bulk RNAseq datasets

Analyses were performed in Galaxy. FASTQ files were groomed using FASTQ Groomer and aligned to the dm6 assembly of the D. melanogaster genome using TopHat. 76-90% of reads for each library mapped to the fly genome, with 3-22% of mapped reads aligning multiple times. We used Cufflinks to quantify the amount of transcription from each gene and assembled a table of FPKMs. We matched gene names to the Refseq “name2” annotations and provide additional gene-level information in the table. Enrichments of canonical markers of KCs (ey) and PNs (Oaz) are shown in Table S5, as well as the pan-neuronal marker nsyb, and pros, which is enriched in mature neurons.

Analysis of single cell RNA sequencing

Analysis of residual transcriptional variation among Kenyon cells of the same type and stage

To explore any residual transcriptional heterogeneity within Kenyon cells of the same class, we first subsetted (subset function in Seurat) individual Kenyon cell clusters at each stage from the integrated clustered data shown in Figure 4E. In each Kenyon cell type, variable genes were identified, followed by scaling of the data and principal component anlaysis (FindVariableFeatures, ScaleData and RunPCA functions). The gene sets belonging to each PC were then plotted over the PC dimensions (FeaturePlot). Results from γm 36h analysis are shown in Figure S 4-2

Transcriptional analysis of cell surface and secreted molecules in Kenyon cells

To compare expression of CSS’s in olfactory Kenyon cells to visual Kenyon cells and non-Kenyon cell brain cells (Figure 5A), log-normalized read counts for each CSS in every cell were obtained by the Seurat function GetAssayData(slot="data", assays ="RNA"). The counts were then summed up for all the CSS’s to get a single value for each cell. This was done to take into account any difference in total number of reads across the 3 types of cell populations.

Expression of Ig-SF and LRR molecules

Expression was measured by computing average expression of each gene in the Kenyon cell population and non-Kenyon cell populations using Seurat’s AverageExpression function (layer="data", assays ="RNA"). The resulted values are in a non-log space, therefore the expression was then normalized by log2 in the heatmaps displayed in Figure S5D.

We used UCell package ⁸⁴ to calculate the gene set enrichment scores at single-cell level for the following groups of genes: Ig-SF+LRR, Ig-SF, LRR, TF, and GPCR. Gene lists are included in the supplemental data. The ranking in UCell uses the Mann-Whitney U statistic.

Re-analysis of published single cell RNA sequencing data

olfactory projection neurons

ScRNA-sequencing data for olfactory projection neurons featured in Figure 7 was acquired from Xie and colleagues ⁵¹. We performed differential expression analysis (FindMarkers function in Seurat, average log2-fold change > 5, adjusted p-value < 0.05) between MZ19-labeled PNs and other PNs to identify molecules uniquely and highly expressed in MZ19⁺ PNs across developmental time. Mz19 PNs were selected as clusters labeled DA1, VA1d, and DC3.

fruitless neurons

The analysis of fruitless neurons was based on the publicly available dataset generated by the Arbeitman lab ⁸⁵: fru+ neurons were purified at 48h APF and subjected to scRNAseq profiling using the 10X Genomics platform. We downloaded the processed dataset from NCBI GEO (GSE160370) and performed a re-analysis of the dataset starting from raw expression counts. The count matrix and metadata were extracted from the Seurat-object for the full dataset, including information on replicates, sex, percentage of mitochondrial transcripts, and cell type annotations. We cross-referenced different names for the same genes to build a consensus set of features for downstream analyses.

The reanalysis was carried out using Seurat v5.0.1 ¹¹⁵. We kept 24,478 cells with more than 1000 transcripts. The preprocessing and clustering were performed using the standard Seurat workflow: gene counts were normalized (function: NormalizeData); 2000 highly-variable genes (FindVariableFeatures) were scaled and used for PCA (functions: ScaleData/RunPCA); first 100 principal components were used for clustering (functions: FindNeighbors/FindClusters, resolution = 2) and tSNE projections (function: RunTSNE).

The initial analysis revealed an unexpected heterogeneity within many clusters that correspond to known cell types (e.g., KCs and visual system neurons). It was driven by activity-regulated genes (ARG), heat-shock proteins (HSP), and components of mitochondrial oxidation phosphorylation complexes (OXPHOS). These signatures are likely to represent unknown sources of biological and technical variability in this dataset (e.g., developmental pseudotime). We used UCell package ⁸⁴ to calculate the gene set enrichment scores for the following groups of genes: ARG (Hr38, sr, CG14186), HSP (Hsp67Ba, Hsp67Bc), OXPHOS (ATPsynbeta, ATPsyngamma). We regressed out these UCell scores together with the numbers of transcripts per cell and replicates (function: ScaleData) and repeated PCA and clustering analysis as described above.

The analysis revealed 132 clusters of neurons. The large transcriptionally distinct clusters are expected to represent abundant cell types with dozens to hundreds of copies per brain. The most abundant fru+ cell types include γ KCs and cell types in the visual system. KC clusters were annotated based on the expression of known marker genes (Figure S7A). The visual system cell types were annotated using a published atlas of the visual system ^54,70. Normalized expression values were averaged for each cluster in fru+ dataset and visual system neurons at 48h APF (atlas V1.1; https://zenodo.org/records/8111612). Next, we identified markers genes enriched in each fru+ cluster (function: FindAllMarkers, adjusted p-value < 0.01, fold-change > 8, detected in more than 50% of cells in the cluster). We used log-transformed expression values of the union of the top 10 marker genes (400 genes) to calculate Pearson’s correlation coefficients between clusters across datasets. The clusters with the best mutual match between datasets and Pearson’s r > 0.9 were considered the same cell types and used for label transfer (Figure S6). We also verified matched clusters based on the expression patterns of marker genes (Figure S7). In total, we annotated 3 KC clusters and 23 visual system clusters. Of those, 7 clusters are matched to neurons in the connectome (Dm3, Tm9, LC10a, LPC1, LLPC1, LLPC2, LLPC3); other clusters with the prefix “N” are yet to be mapped to morphological cell types. Clusters with the prefix “X” did not match clusters in the visual system atlas and we infer these are non-optic-lobe types. While the Palmateer dataset includes the ventral nerve cord, individual fruitless neuron types from the VNC are relatively rare, thus larger "X" clusters likely derive from the central brain.

Image analysis

Analysis considerations

We used mixed-sex populations. Sex differences in the fly are well-documented, and anatomic and physiologic sex differences have not been observed in the mushroom body. Any brains that appeared damaged from dissections, or those with the mushroom body region obscured due to insufficient tracheal removal, were not included in the analysis.

Analysis was performed blind to genotype, and also blind to the goals of the experiment when possible, and quantitation of features on the anterior and posterior sides of the brain were recorded independent of one another and merged after all quantifications were completed.

Area of the largest slice of the calyx

To measure the size of the mushroom body calyx, we used markers such as ChAT or brp to visualize the structure. We then measured the area of each slice of the calyx in Z by outlining in FIJI and using the ‘Measure’ command. The area of the largest slice of the calyx was reported. Previously, we found this correlated well with the number of Kenyon cells innervating the calyx ²¹. We were only able to make this measurement in brains mounted with 90% glycerol and propyl gallate. Brains mounted in DPX were more variable in size in different batches and were substantially different in size from those mounted in glycerol and propyl gallate.

Projection neuron bouton counts

To count aggregate boutons, we used ChAT signal as previously described ²¹. We counted as separate structures ChAT signals that were compact and appeared distinct from one another and that were 2+ micrometers in diameter. We found that boutons in slice 0 often appeared in slices −1 and +1 as well, but never in slices −2 or +2. To avoid over counting, we began at the most superficial slice in the stack where boutons were visible, and counted every other slice, i.e. every second micron.

To count boutons of subsets of PNs, we drove a fluorescent reporter under the control of VT033008, VT033006, or MZ19 and counted coherent and compact fluorescent signals, that overlapped with ChAT signals in order to distinguish true signal from background noise.

Projection neuron cell counts

To count PNs, we used genetically-encoded fluorescence in that group of cells. We counted labeled somata in every third slice in the stack (every third micron along the A-P axis), with reference to DAPI to distinguish individual cells from one another. As we did for boutons, in analyzing somata we initially determined that somata in slice 0 could also be seen in slices −2, −1, +1, and +2 but not in slice −3 or +3. To avoid double-counting, we therefore counted every third micron.

Normalized Bruchpilot and Discs large intensity

We measured Bruchpilot (Brp) and Discs Large (Dlg) intensity in the calyx as a readout of synaptic density. First, we identified the Z plane with the largest extent of the calyx in the A-P axis, and measured the mean flourescence value in a defined ROI region of 250² pixels. The defined ROI was placed in the center of the calyx. For normalization, the calyx Brp signal was divided by Brp signal measured in a 75² pixel ROI in an unmanipulated brain region, the protocerebral bridge. We chose this brain region as it was the closest to the calyx and was in the field of view in all calyx images taken.

Normalized Kenyon cell fluorescence in bouton ROIs

We first defined ROIs in Fiji for each projection neuron bouton labeled by MZ19. Boutons were defined by using genetically-encoded fluorescence in that group of cells, overlapped with ChAT signal to distinguish true bouton signal from any background noise. In line with PN bouton counting methods described earlier, we only made bouton ROIs in every second slice (z step size was 1 micron). As a control, we also made bouton ROIs not labeled by MZ19. These non-MZ19 boutons were nearby MZ19+ boutons. We also matched the number of non-MZ19 bouton ROIs to the number of MZ19+ bouton ROIs for each z-slice of the calyx. For each bouton ROI, we then measured the mean fluorescence value in the Kenyon cell channel and normalized it to the mean fluorescence in the Kenyon cell channel in the calyx ROI defined by ChAT immunoreactivity. When measuring Dlg intensity in bouton ROIs we used this same procedure but only for MZ19+ boutons as we could not make ROIs for unlabeled boutons. Bouton ROI signal was normalized to the mean fluorescence value of Dlg in the calyx ROI defined by Dlg signal.

Kenyon cell claw counts

To count claws on single Kenyon cells we used images of single Kenyon cells filled with Texas red dextran. Claws were identified as concentrations of membrane connected to the brightly labeled Kenyon cell axon. Images were acquired with 1 micron steps and individual claws could be visible in multiple stacks but were well spaced and thus easily discernable from one another.

Statistical considerations

Brains were prepared for imaging in batches of 5–10. In initial batches, we assessed the variability of the manipulation, and used this variability to determine how many batches to analyze so as to obtain enough informative samples. Genotypes or conditions being compared with one another were always prepared for staining together and imaged interspersed with one another to equalize batch effects. We excluded from analysis samples with overt physical damage to the cells or structures being measured.

Statistical tests applied are mentioned in each figure legend. P-value is written in the figure above the relevant comparison. Normality of data was assessed with Shapiro test. When both distributions were normal (p>0.05) we used two-sided Student’s T-test. When one or both distributions were non-normal (p<=0.05) we used the two-sided Wilcoxon rank sum test. To communicate our findings in the simplest and most complete way, we have displayed each data point for each sample to allow readers to assess effect size and significance directly. Table S1 includes the n, name of statistical test applied, test value, p-value and Cohen’s D effect size calculated using a pooled standard deviation. Statistical analyses were performed in R and analysis code can be found on GitHub (see data availability).

Supplementary Material

Supplementary Information

NIHMS2125319-supplement-Supplementary_Information.pdf^{(2.3MB, pdf)}

Acknowledgements

We thank Bloomington Drosophila Stock Center, Liqun Luo, Yoshi Aso, and Gerry Rubin and Heather Dionne for sharing fly strains; and Kamlai Saiya-Cork (University of Michigan Flow Cytometry Core) and Gregg Sobocinski (MCDB Imaging Core) for technical assistance. Single cell processing and next-generation sequencing was carried out in the Advanced Genomics Core at the University of Michigan. Margarita Brovkina and Audrey Drotos provided conceptual insight. Vanessa Ruta and Claude Desplan provided support for the project, and Ruta lab members assisted in generation of bulk sequencing libraries. Robin Hiesinger and Clowney Lab members provided input on the manuscript. EMT-K was supported by NIH T32 DC00011, MA by the University of Michigan Rackham Predoctoral Fellowship and Rackham Graduate Student Research Grant, and FRG by the MNI Magnificent Michigan Fellowship. B.S. was supported by a Long-Term Fellowship LT000010/2020-L from the Human Frontier Science Program. EJC is a McKnight Scholar, Pew Biomedical Scholar, and Rita Allen Milton Cassell Scholar. Funding was provided by NIDCD R01DC018032 (to EJC)

Footnotes

Declaration of Interests

The authors declare no competing interests.

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, E. Josephine Clowney (jclowney@umich.edu).

Materials Availability

This study did not generate new unique reagents.

Data and Code Availability

We used data from the Hemibrain connectome that are available to the public under a Creative Commons license at https://neuprint.janelia.org. Annotated tables of synapses, projection neurons, projection neuron boutons and branch points, Kenyon cells, and Kenyon cell claws are can be found at 10.5281/zenodo.15425075. Bulk RNA sequencing of Kenyon cells and projection neurons is available under NCBI GEO accession GSE276028. scRNAseq of Kenyon cells at developmental time points is available under NCBI GEO accession GSE296723; this includes metadata for individual single cells, single-cell gene expression matrix and UMAP embeddings. Metadata for single cells includes sample identities (dataset, developmental stage, genotype/driver information) and cluster identities (classes, types, subtypes). scRNAseq of adult Kenyon cells is available under NCBI GEO accession GSE274592. Re-analyzed fruitless neurons from Palmateer et al.⁸⁵ is available at 10.5281/zenodo.15467833.
Code for Hemibrain analyses is available at 10.5281/zenodo.15425075. Code for reanalysis of fruitless neurons from Palmateer et al.⁸⁵ is available at 10.5281/zenodo.15467864
Any additional information required to reanalyze the data reported in this work is available from the Lead Contact upon request.

References

1.Marr D. (1969). A theory of cerebellar cortex. J. Physiol 202, 437–470. 10.1113/jphysiol.1969.sp008820. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Albus JS (1971). A theory of cerebellar function. Math. Biosci 10, 25–61. 10.1016/0025-5564(71)90051-4. [DOI] [Google Scholar]
3.Ito M. (1970). Neurophysiological aspects of the cerebellar motor control system. Int. J. Neurol 7, 162–176. [PubMed] [Google Scholar]
4.Chabrol FP, Arenz A, Wiechert MT, Margrie TW, and DiGregorio DA (2015). Synaptic diversity enables temporal coding of coincident multisensory inputs in single neurons. Nat. Neurosci 18, 718–727. 10.1038/nn.3974. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Huang C-C, Sugino K, Shima Y, Guo C, Bai S, Mensh BD, Nelson SB, and Hantman AW (2013). Convergence of pontine and proprioceptive streams onto multimodal cerebellar granule cells. Elife 2, e00400. 10.7554/eLife.00400. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Gruntman E, and Turner GC (2013). Integration of the olfactory code across dendritic claws of single mushroom body neurons. Nat. Neurosci 16, 1821–1829. 10.1038/nn.3547. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Ishikawa T, Shimuta M, and Häusser M. (2015). Multimodal sensory integration in single cerebellar granule cells in vivo. Elife 4. 10.7554/eLife.12916. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Cayco-Gajic NA, Clopath C, and Silver RA (2017). Sparse synaptic connectivity is required for decorrelation and pattern separation in feedforward networks. Nat. Commun 8, 1116. 10.1038/s41467-017-01109-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Litwin-Kumar A, Harris KD, Axel R, Sompolinsky H, and Abbott LF (2017). Optimal Degrees of Synaptic Connectivity. Neuron 93, 1153–1164.e7. 10.1016/j.neuron.2017.01.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Cayco-Gajic NA, and Silver RA (2019). Re-evaluating Circuit Mechanisms Underlying Pattern Separation. Neuron 101, 584–602. 10.1016/j.neuron.2019.01.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Turner GC, Bazhenov M, and Laurent G. (2008). Olfactory representations by Drosophila mushroom body neurons. J. Neurophysiol 99, 734–746. 10.1152/jn.01283.2007. [DOI] [PubMed] [Google Scholar]
12.Farris SM (2011). Are mushroom bodies cerebellum-like structures? Arthropod Struct. Dev 40, 368–379. 10.1016/j.asd.2011.02.004. [DOI] [PubMed] [Google Scholar]
13.Sawtell NB (2010). Multimodal integration in granule cells as a basis for associative plasticity and sensory prediction in a cerebellum-like circuit. Neuron 66, 573–584. 10.1016/j.neuron.2010.04.018. [DOI] [PubMed] [Google Scholar]
14.Hobbs MJ, and Young JZ (1973). A cephalopod cerebellum. Brain Res. 55, 424–430. 10.1016/0006-8993(73)90307-7. [DOI] [PubMed] [Google Scholar]
15.Leutgeb JK, Leutgeb S, Moser M-B, and Moser EI (2007). Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science 315, 961–966. 10.1126/science.1135801. [DOI] [PubMed] [Google Scholar]
16.Srinivasan S, and Stevens CF (2018). The distributed circuit within the piriform cortex makes odor discrimination robust. J. Comp. Neurol 526, 2725–2743. 10.1002/cne.24492. [DOI] [PubMed] [Google Scholar]
17.Aso Y, Grübel K, Busch S, Friedrich AB, Siwanowicz I, and Tanimoto H. (2009). The mushroom body of adult Drosophila characterized by GAL4 drivers. J. Neurogenet 23, 156–172. 10.1080/01677060802471718. [DOI] [PubMed] [Google Scholar]
18.Caron SJC, Ruta V, Abbott LF, and Axel R. (2013). Random convergence of olfactory inputs in the Drosophila mushroom body. Nature 497, 113–117. 10.1038/nature12063. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Jefferis GS, Marin EC, Stocker RF, and Luo L. (2001). Target neuron prespecification in the olfactory map of Drosophila. Nature 414, 204–208. 10.1038/35102574. [DOI] [PubMed] [Google Scholar]
20.Ahmed M, Rajagopalan AE, Pan Y, Li Y, Williams DL, Pedersen EA, Thakral M, Previero A, Close KC, Christoforou CP, et al. (2023). Input density tunes Kenyon cell sensory responses in the Drosophila mushroom body. Curr. Biol 33, 2742–2760.e12. 10.1016/j.cub.2023.05.064. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Elkahlah NA, Rogow JA, Ahmed M, and Clowney EJ (2020). Presynaptic developmental plasticity allows robust sparse wiring of the drosophila mushroom body. Elife 9. 10.7554/eLife.52278. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Eichler K, Li F, Litwin-Kumar A, Park Y, Andrade I, Schneider-Mizell CM, Saumweber T, Huser A, Eschbach C, Gerber B, et al. (2017). The complete connectome of a learning and memory centre in an insect brain. Nature 548, 175–182. 10.1038/nature23455. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Li F, Lindsey JW, Marin EC, Otto N, Dreher M, Dempsey G, Stark I, Bates AS, Pleijzier MW, Schlegel P, et al. (2020). The connectome of the adult Drosophila mushroom body provides insights into function. Elife 9. 10.7554/eLife.62576. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Murthy M, Fiete I, and Laurent G. (2008). Testing odor response stereotypy in the Drosophila mushroom body. Neuron 59, 1009–1023. 10.1016/j.neuron.2008.07.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Zheng Z, Li F, Fisher C, Ali IJ, Sharifi N, Calle-Schuler S, Hsu J, Masoodpanah N, Kmecova L, Kazimiers T, et al. (2022). Structured sampling of olfactory input by the fly mushroom body. Curr. Biol 32, 3334–3349.e6. 10.1016/j.cub.2022.06.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Ito K, Awano W, Suzuki K, Hiromi Y, and Yamamoto D. (1997). The Drosophila mushroom body is a quadruple structure of clonal units each of which contains a virtually identical set of neurones and glial cells. Development 124, 761–771. 10.1242/dev.124.4.761. [DOI] [PubMed] [Google Scholar]
27.Kunz T, Kraft KF, Technau GM, and Urbach R. (2012). Origin of Drosophila mushroom body neuroblasts and generation of divergent embryonic lineages. Development 139, 2510–2522. 10.1242/dev.077883. [DOI] [PubMed] [Google Scholar]
28.Lee T, Lee A, and Luo L. (1999). Development of the Drosophila mushroom bodies: Sequential generation of three distinct types of neurons from a neuroblast. Development 126, 4065–4076. [DOI] [PubMed] [Google Scholar]
29.Yu H-H, Kao C-F, He Y, Ding P, Kao J-C, and Lee T. (2010). A complete developmental sequence of a Drosophila neuronal lineage as revealed by twin-spot MARCM. PLoS Biol. 8, e1000461. 10.1371/journal.pbio.1000461. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Lin S, Kao C-F, Yu H-H, Huang Y, and Lee T. (2012). Lineage analysis of Drosophila lateral antennal lobe neurons reveals notch-dependent binary temporal fate decisions. PLoS Biol. 10, e1001425. 10.1371/journal.pbio.1001425. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Ganguly I, Heckman EL, Litwin-Kumar A, Clowney EJ, and Behnia R. (2024). Diversity of visual inputs to Kenyon cells of the Drosophila mushroom body. Nat. Commun 15, 5698. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Wong AM, Wang JW, and Axel R. (2002). Spatial representation of the glomerular map in the Drosophila protocerebrum. Cell 109, 229–241. 10.1016/s0092-8674(02)00707-9. [DOI] [PubMed] [Google Scholar]
33.Marin EC, Jefferis GSXE, Komiyama T, Zhu H, and Luo L. (2002). Representation of the glomerular olfactory map in the Drosophila brain. Cell 109, 243–255. 10.1016/S0092-8674(02)00700-6. [DOI] [PubMed] [Google Scholar]
34.Tanaka NK, Awasaki T, Shimada T, and Ito K. (2004). Integration of chemosensory pathways in the Drosophila second-order olfactory centers. Curr. Biol 14, 449–457. 10.1016/j.cub.2004.03.006. [DOI] [PubMed] [Google Scholar]
35.Jefferis GSXE, Vyas RM, Berdnik D, Ramaekers A, Stocker RF, Tanaka NK, Ito K, and Luo L. (2004). Developmental origin of wiring specificity in the olfactory system of Drosophila. Development 131, 117–130. 10.1242/dev.00896. [DOI] [PubMed] [Google Scholar]
36.Yang J-Y, O’Connell TF, Hsu W-MM, Bauer MS, Dylla KV, Sharpee TO, and Hong EJ (2023). Restructuring of olfactory representations in the fly brain around odor relationships in natural sources. bioRxiv, 2023.02.15.528627. 10.1101/2023.02.15.528627. [DOI] [Google Scholar]
37.Hayashi TT, MacKenzie AJ, Ganguly I, Ellis KE, Smihula HM, Jacob MS, Litwin-Kumar A, and Caron SJC (2022). Mushroom body input connections form independently of sensory activity in Drosophila melanogaster. Curr. Biol 0. 10.1016/j.cub.2022.07.055. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Ellis KE, Bervoets S, Smihula H, Ganguly I, Vigato E, Auer TO, Benton R, Litwin-Kumar A, and Caron SJC (2024). Evolution of connectivity architecture in the Drosophila mushroom body. Nat. Commun 15, 4872. 10.1038/s41467-024-48839-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Lin H-H, Lai JS-Y, Chin A-L, Chen Y-C, and Chiang A-S (2007). A map of olfactory representation in the Drosophila mushroom body. Cell 128, 1205–1217. 10.1016/j.cell.2007.03.006. [DOI] [PubMed] [Google Scholar]
40.Campbell RAA, Honegger KS, Qin H, Li W, Demir E, and Turner GC (2013). Imaging a population code for odor identity in the Drosophila mushroom body. J. Neurosci 33, 10568–10581. 10.1523/JNEUROSCI.0682-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Bengtsson F, and Jörntell H. (2009). Sensory transmission in cerebellar granule cells relies on similarly coded mossy fiber inputs. Proc. Natl. Acad. Sci. U. S. A 106, 2389–2394. 10.1073/pnas.0808428106. [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Shuster SA, Wagner MJ, Pan-Doh N, Ren J, Grutzner SM, Beier KT, Kim TH, Schnitzer MJ, and Luo L. (2021). The relationship between birth timing, circuit wiring, and physiological response properties of cerebellar granule cells. Proc. Natl. Acad. Sci. U. S. A 118. 10.1073/pnas.2101826118. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Kennedy A, Wayne G, Kaifosh P, Alviña K, Abbott LF, and Sawtell NB (2014). A temporal basis for predicting the sensory consequences of motor commands in an electric fish. Nat. Neurosci 17, 416–422. 10.1038/nn.3650. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Gilmer JI, and Person AL (2017). Morphological Constraints on Cerebellar Granule Cell Combinatorial Diversity. J. Neurosci 37, 12153–12166. 10.1523/JNEUROSCI.0588-17.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Nguyen TM, Thomas LA, Rhoades JL, Ricchi I, Yuan XC, Sheridan A, Hildebrand DGC, Funke J, Regehr WG, and Lee W-CA (2023). Structured cerebellar connectivity supports resilient pattern separation. Nature 613, 543–549. 10.1038/s41586-022-05471-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Scheffer LK, Xu CS, Januszewski M, Lu Z, Takemura S-Y, Hayworth KJ, Huang GB, Shinomiya K, Maitlin-Shepard J, Berg S, et al. (2020). A connectome and analysis of the adult Drosophila central brain. Elife 9. 10.7554/eLife.57443. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Zhu S, Chiang A-S, and Lee T. (2003). Development of the Drosophila mushroom bodies: elaboration, remodeling and spatial organization of dendrites in the calyx. Development 130, 2603–2610. 10.1242/dev.00466. [DOI] [PubMed] [Google Scholar]
48.Jefferis GSXE, Potter CJ, Chan AM, Marin EC, Rohlfing T, Maurer CR Jr, and Luo L. (2007). Comprehensive maps of Drosophila higher olfactory centers: spatially segregated fruit and pheromone representation. Cell 128, 1187–1203. 10.1016/j.cell.2007.01.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Zheng Z, Lauritzen JS, Perlman E, Robinson CG, Nichols M, Milkie D, Torrens O, Price J, Fisher CB, Sharifi N, et al. (2018). A Complete Electron Microscopy Volume of the Brain of Adult Drosophila melanogaster. Cell 174, 730–743.e22. 10.1016/j.cell.2018.06.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Li H, Horns F, Xie Q, Xie Q, Li T, Luginbuhl DJ, Luo L, and Quake SR (2017). Classifying Drosophila Olfactory Projection Neuron Subtypes by Single-Cell RNA Sequencing. Cell 171, 1206.e22–1220.e22. 10.1016/j.cell.2017.10.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Xie Q, Brbic M, Horns F, Kolluru SS, Jones RC, Li J, Reddy AR, Xie A, Kohani S, Li Z, et al. (2021). Temporal evolution of single-cell transcriptomes of Drosophila olfactory projection neurons. Elife 10. 10.7554/eLife.63450. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Carrillo RA, Özkan E, Menon KP, Nagarkar-Jaiswal S, Lee PT, Jeon M, Birnbaum ME, Bellen HJ, Garcia KC, and Zinn K. (2015). Control of Synaptic Connectivity by a Network of Drosophila IgSF Cell Surface Proteins. Cell 163, 1770–1782. 10.1016/j.cell.2015.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Xu C, Theisen E, Maloney R, Peng J, Santiago I, Yapp C, Werkhoven Z, Rumbaut E, Shum B, Tarnogorska D, et al. (2019). Control of Synaptic Specificity by Establishing a Relative Preference for Synaptic Partners. Neuron 103, 865–877.e7. 10.1016/j.neuron.2019.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Yoo J, Dombrovski M, Mirshahidi P, Nern A, LoCascio SA, Zipursky SL, and Kurmangaliyev YZ (2023). Brain wiring determinants uncovered by integrating connectomes and transcriptomes. Curr. Biol 33, 3998–4005.e6. 10.1016/j.cub.2023.08.020. [DOI] [PubMed] [Google Scholar]
55.Özkan E, Carrillo RA, Eastman CL, Weiszmann R, Waghray D, Johnson KG, Zinn K, Celniker SE, and Garcia KC (2013). An extracellular interactome of immunoglobulin and LRR proteins reveals receptor-ligand networks. Cell 154, 228. 10.1016/j.cell.2013.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Tan L, Zhang KX, Pecot MY, Nagarkar-Jaiswal S, Lee PT, Takemura SY, McEwen JM, Nern A, Xu S, Tadros W, et al. (2015). Ig Superfamily Ligand and Receptor Pairs Expressed in Synaptic Partners in Drosophila. Cell 163, 1756–1769. 10.1016/j.cell.2015.11.021.Ig. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Bates AS, Schlegel P, Roberts RJV, Drummond N, Tamimi IFM, Turnbull R, Zhao X, Marin EC, Popovici PD, Dhawan S, et al. (2020). Complete Connectomic Reconstruction of Olfactory Projection Neurons in the Fly Brain. Curr. Biol 30, 3183–3199.e6. 10.1016/j.cub.2020.06.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Yusuyama K, Meinertzhagen IA, and Schürmann FW (2002). Synaptic organization of the mushroom body calyx in Drosophila melanogaster. J. Comp. Neurol 445, 211–226. 10.1002/cne.10155. [DOI] [PubMed] [Google Scholar]
59.Leiss F, Groh C, Butcher NJ, Meinertzhagen IA, and Tavosanis G. (2009). Synaptic organization in the adult Drosophila mushroom body calyx. J. Comp. Neurol 517, 808–824. 10.1002/cne.22184. [DOI] [PubMed] [Google Scholar]
60.Yang K, Liu T, Wang Z, Liu J, Shen Y, Pan X, Wen R, Xie H, Ruan Z, Tan Z, et al. (2022). Classifying Drosophila olfactory projection neuron boutons by quantitative analysis of electron microscopic reconstruction. iScience 25, 104180. 10.1016/j.isci.2022.104180. [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Butcher NJ, Friedrich AB, Lu Z, Tanimoto H, and Meinertzhagen IA (2012). Different classes of input and output neurons reveal new features in microglomeruli of the adult Drosophila mushroom body calyx. J. Comp. Neurol 520, 2185–2201. 10.1002/cne.23037. [DOI] [PubMed] [Google Scholar]
62.Kurusu M, Cording A, Taniguchi M, Menon K, Suzuki E, and Zinn K. (2008). A screen of cell-surface molecules identifies leucine-rich repeat proteins as key mediators of synaptic target selection. Neuron 59, 972–985. 10.1016/j.neuron.2008.07.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Muthukumar AK, Stork T, and Freeman MR (2014). Activity-dependent regulation of astrocyte GAT levels during synaptogenesis. Nat. Neurosci 17, 1340–1350. 10.1038/nn.3791. [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Croset V, Treiber CD, and Waddell S. (2018). Cellular diversity in the Drosophila midbrain revealed by single-cell transcriptomics. Elife 7, 1–31. 10.7554/eLife.34550. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Lee G, Foss M, Goodwin SF, Carlo T, Taylor BJ, and Hall JC (2000). Spatial, temporal, and sexually dimorphic expression patterns of the fruitless gene in the Drosophila central nervous system. Journal of Neurobiology 43, 404–426. . [DOI] [PubMed] [Google Scholar]
66.Shih MFM, Davis FP, Henry GL, and Dubnau J. (2019). Nuclear transcriptomes of the seven neuronal cell types that constitute the drosophila mushroom bodies. G3: Genes, Genomes, Genetics 9, 81–94. 10.1534/g3.118.200726. [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Vogt K, Aso Y, Hige T, Knapek S, Ichinose T, Friedrich AB, Turner GC, Rubin GM, and Tanimoto H. (2016). Direct neural pathways convey distinct visual information to drosophila mushroom bodies. Elife 5, 1–13. 10.7554/eLife.14009. [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Li J, Mahoney BD, Jacob MS, and Caron SJC (2020). Visual Input into The Drosophila Melanogaster Mushroom Body. SSRN Electronic Journal, 1–47. 10.2139/ssrn.3595714. [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Jain S, Lin Y, Kurmangaliyev YZ, Valdes-Aleman J, LoCascio SA, Mirshahidi P, Parrington B, and Zipursky SL (2022). A global timing mechanism regulates cell-type-specific wiring programmes. Nature 603, 112–118. 10.1038/s41586-022-04418-5. [DOI] [PubMed] [Google Scholar]
70.Kurmangaliyev YZ, Yoo J, Valdes-Aleman J, Sanfilippo P, and Zipursky SL (2020). Transcriptional Programs of Circuit Assembly in the Drosophila Visual System. Neuron 108, 1045–1057.e6. 10.1016/j.neuron.2020.10.006. [DOI] [PubMed] [Google Scholar]
71.Alyagor I, Berkun V, Keren-Shaul H, Marmor-Kollet N, David E, Mayseless O, Issman-Zecharya N, Amit I, and Schuldiner O. (2018). Combining developmental and perturbation-seq uncovers transcriptional modules orchestrating neuronal remodeling. Dev. Cell 47, 38–52.e6. 10.1016/j.devcel.2018.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Ahmed M. (2023). Generation of sparse, combinatorial wiring for sensory coding. 10.7302/8559. [DOI] [Google Scholar]
73.Li H, Watson A, Olechwier A, Anaya M, Sorooshyari SK, Harnett DP, Lee H-KP, Vielmetter J, Fares MA, Garcia KC, et al. (2017). Deconstruction of the beaten Path-Sidestep interaction network provides insights into neuromuscular system development. Elife 6. 10.7554/eLife.28111. [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Barish S, Nuss S, Strunilin I, Bao S, Mukherjee S, Jones CD, and Volkan PC (2018). Combinations of DIPs and Dprs control organization of olfactory receptor neuron terminals in Drosophila. PLOS Genetics 14, 1–33. 10.1371/journal.pgen.1007560. [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Brovero SG, Fortier JC, Hu H, Lovejoy PC, Newell NR, Palmateer CM, Tzeng R-Y, Lee P-T, Zinn K, and Arbeitman MN (2021). Investigation of Drosophila fruitless neurons that express Dpr/DIP cell adhesion molecules. Elife 10. 10.7554/eLife.63101. [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Sanes JR, and Zipursky SL (2020). Synaptic Specificity, Recognition Molecules, and Assembly of Neural Circuits. Cell 181, 536–556. 10.1016/j.cell.2020.04.008. [DOI] [PubMed] [Google Scholar]
77.Hiesinger PR (2021). Brain wiring with composite instructions. Bioessays 43, e2000166. 10.1002/bies.202000166. [DOI] [PubMed] [Google Scholar]
78.Hassan BA, and Hiesinger PR (2015). Beyond Molecular Codes: Simple Rules to Wire Complex Brains. Cell 163, 285–291. 10.1016/j.cell.2015.09.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Menon KP, Kulkarni V, Takemura S-Y, Anaya M, and Zinn K. (2019). Interactions between Dpr11 and DIP-γ control selection of amacrine neurons in Drosophila color vision circuits. Elife 8. 10.7554/eLife.48935. [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Bornstein B, Meltzer H, Adler R, Alyagor I, Berkun V, Cummings G, Reh F, Keren-Shaul H, David E, Riemensperger T, et al. (2021). Transneuronal Dpr12/DIP-δ interactions facilitate compartmentalized dopaminergic innervation of Drosophila mushroom body axons. EMBO J. 40, e105763. 10.15252/embj.2020105763. [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Courgeon M, and Desplan C. (2019). Coordination between stochastic and deterministic specification in the Drosophila visual system. Science 366, eaay6727. 10.1126/science.aay6727. [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Cheng S, Park Y, Kurleto JD, Jeon M, Zinn K, Thornton JW, and Özkan E. (2019). Family of neural wiring receptors in bilaterians defined by phylogenetic, biochemical, and structural evidence. Proc. Natl. Acad. Sci. U. S. A 116, 9837–9842. 10.1073/pnas.1818631116. [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Morano NC, Lopez DH, Meltzer H, Sergeeva AP, Katsamba PS, Rostam KD, Gupta HP, Becker JE, Bornstein B, Cosmanescu F, et al. (2025). Members of the DIP and Dpr adhesion protein families use cis inhibition to shape neural development in Drosophila. PLoS Biol. 23, e3003030. 10.1371/journal.pbio.3003030. [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Andreatta M, and Carmona SJ (2021). UCell: Robust and scalable single-cell gene signature scoring. Comput. Struct. Biotechnol. J 19, 3796–3798. 10.1016/j.csbj.2021.06.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Palmateer CM, Artikis C, Brovero SG, Friedman B, Gresham A, and Arbeitman MN (2023). Single-cell transcriptome profiles of Drosophila fruitless-expressing neurons from both sexes. Elife 12. 10.7554/eLife.78511. [DOI] [PMC free article] [PubMed] [Google Scholar]
86.Özel MN, Simon F, Jafari S, Holguera I, Chen Y-C, Benhra N, El-Danaf RN, Kapuralin K, Malin JA, Konstantinides N, et al. (2021). Neuronal diversity and convergence in a visual system developmental atlas. Nature 589, 88–95. 10.1038/s41586-020-2879-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
87.Yagi R, Mabuchi Y, Mizunami M, and Tanaka NK (2016). Convergence of multimodal sensory pathways to the mushroom body calyx in Drosophila melanogaster. Sci. Rep 6, 1–8. 10.1038/srep29481. [DOI] [PMC free article] [PubMed] [Google Scholar]
88.Rosenblatt F. (1958). THE PERCEPTRON: A PROBABILISTIC MODEL FOR INFORMATION STORAGE AND ORGANIZATION IN THE BRAIN. Psychol. Rev 65, 386–408. [DOI] [PubMed] [Google Scholar]
89.De Belle JS, and Heisenberg M. (1994). Associative odor learning in Drosophila abolished by chemical ablation of mushroom bodies. Science 263, 692–695. 10.1126/science.8303280. [DOI] [PubMed] [Google Scholar]
90.Masse NY, Turner GC, and Jefferis GSXE (2009). Olfactory Information Processing in Drosophila. Curr. Biol 19, R700–R713. 10.1016/j.cub.2009.06.026. [DOI] [PubMed] [Google Scholar]
91.Laurent G, and Naraghi M. (1994). Odorant-induced oscillations in the mushroom bodies of the locust. J. Neurosci 14, 2993–3004. 10.1523/jneurosci.14-05-02993.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
92.Di Bella DJ, Habibi E, Stickels RR, Scalia G, Brown J, Yadollahpour P, Yang SM, Abbate C, Biancalani T, Macosko EZ, et al. (2021). Molecular logic of cellular diversification in the mouse cerebral cortex. Nature 595, 554–559. 10.1038/s41586-021-03670-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
93.Chen Y, Chen X, Baserdem B, Zhan H, Li Y, Davis MB, Kebschull JM, Zador AM, Koulakov AA, and Albeanu DF (2022). High-throughput sequencing of single neuron projections reveals spatial organization in the olfactory cortex. Cell 185, 4117–4134.e28. 10.1016/j.cell.2022.09.038. [DOI] [PMC free article] [PubMed] [Google Scholar]
94.Yu Y-C, Bultje RS, Wang X, and Shi S-H (2009). Specific synapses develop preferentially among sister excitatory neurons in the neocortex. Nature 458, 501–504. 10.1038/nature07722. [DOI] [PMC free article] [PubMed] [Google Scholar]
95.Li Y, Lu H, Cheng P-L, Ge S, Xu H, Shi S-H, and Dan Y. (2012). Clonally related visual cortical neurons show similar stimulus feature selectivity. Nature 486, 118–121. 10.1038/nature11110. [DOI] [PMC free article] [PubMed] [Google Scholar]
96.Kasthuri N, Hayworth KJ, Berger DR, Schalek RL, Conchello JA, Knowles-Barley S, Lee D, Vázquez-Reina A, Kaynig V, Jones TR, et al. (2015). Saturated Reconstruction of a Volume of Neocortex. Cell 162, 648–661. 10.1016/j.cell.2015.06.054. [DOI] [PubMed] [Google Scholar]
97.Wolterhoff N, and Hiesinger PR (2024). Synaptic promiscuity in brain development. Curr. Biol 34, R102–R116. 10.1016/j.cub.2023.12.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
98.Bekkers JM, and Stevens CF (1991). Excitatory and inhibitory autaptic currents in isolated hippocampal neurons maintained in cell culture. Proc. Natl. Acad. Sci. U. S. A 88, 7834–7838. 10.1073/pnas.88.17.7834. [DOI] [PMC free article] [PubMed] [Google Scholar]
99.Van der Loos H, and Glaser EM (1972). Autapses in neocortex cerebri: synapses between a pyramidal cell’s axon and its own dendrites. Brain Res. 48, 355–360. 10.1016/0006-8993(72)90189-8. [DOI] [PubMed] [Google Scholar]
100.Grueber WB, and Sagasti A. (2010). Self-avoidance and tiling: Mechanisms of dendrite and axon spacing. Cold Spring Harb. Perspect. Biol 2, a001750. 10.1101/cshperspect.a001750. [DOI] [PMC free article] [PubMed] [Google Scholar]
101.Williams DL, Sikora VM, Hammer MA, Amin S, Brinjikji T, Brumley EK, Burrows CJ, Carrillo PM, Cromer K, Edwards SJ, et al. (2021). May the odds be ever in your favor: Non-deterministic mechanisms diversifying cell surface molecule expression. Front. Cell Dev. Biol 9, 720798. 10.3389/fcell.2021.720798. [DOI] [PMC free article] [PubMed] [Google Scholar]
102.Valdes-Aleman J, Fetter RD, Sales EC, Heckman EL, Venkatasubramanian L, Doe CQ, Landgraf M, Cardona A, and Zlatic M. (2021). Comparative Connectomics Reveals How Partner Identity, Location, and Activity Specify Synaptic Connectivity in Drosophila. Neuron 109, 105–122.e7. 10.1016/j.neuron.2020.10.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
103.Kiral FR, Linneweber GA, Mathejczyk T, Georgiev SV, Wernet MF, Hassan BA, von Kleist M, and Hiesinger PR (2020). Autophagy-dependent filopodial kinetics restrict synaptic partner choice during Drosophila brain wiring. Nat. Commun 11, 1325. 10.1038/s41467-020-14781-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
104.Kiral FR, Dutta SB, Linneweber GA, Hilgert S, Poppa C, Duch C, von Kleist M, Hassan BA, and Hiesinger PR (2021). Brain connectivity inversely scales with developmental temperature in Drosophila. Cell Rep. 37, 110145. 10.1016/j.celrep.2021.110145. [DOI] [PubMed] [Google Scholar]
105.Osaka J, Ishii A, Wang X, Iwanaga R, Kawamura H, Akino S, Sugie A, Hakeda-Suzuki S, and Suzuki T. (2024). Complex formation of immunoglobulin superfamily molecules Side-IV and Beat-IIb regulates synaptic specificity. Cell Rep. 43, 113798. 10.1016/j.celrep.2024.113798. [DOI] [PubMed] [Google Scholar]
106.Drotos AC, and Roberts MT (2024). Identifying neuron types and circuit mechanisms in the auditory midbrain. Hear. Res 442, 108938. 10.1016/j.heares.2023.108938. [DOI] [PMC free article] [PubMed] [Google Scholar]
107.Minsky M. (1952). A Neural-Analogue Calculator Based upon a Probability Model of Reinforcement. Preprint. [Google Scholar]
108.Minsky M. (2016). Building my randomly wired neural network machine. [Google Scholar]
109.Edelman GM (1989). The remembered present: A biological theory of consciousness. The remembered present: A biological theory of consciousness. [Google Scholar]
110.Brovkina MV, Duffié R, Burtis AEC, and Clowney EJ (2021). Fruitless decommissions regulatory elements to implement cell-type-specific neuronal masculinization. PLoS Genet. 17, e1009338. 10.1371/journal.pgen.1009338. [DOI] [PMC free article] [PubMed] [Google Scholar]
111.Aso Y, Hattori D, Yu Y, Johnston RM, Iyer NA, Ngo TTB, Dionne H, Abbott LF, Axel R, Tanimoto H, et al. (2014). The neuronal architecture of the mushroom body provides a logic for associative learning. Elife 3, e04577. 10.7554/eLife.04577. [DOI] [PMC free article] [PubMed] [Google Scholar]
112.Zappia L, and Oshlack A. (2018). Clustering trees: a visualization for evaluating clusterings at multiple resolutions. GigaScience 7, giy083. 10.1093/gigascience/giy083. [DOI] [PMC free article] [PubMed] [Google Scholar]
113.Pei J, Kinch LN, and Grishin NV (2018). FlyXCDB—A Resource for Drosophila Cell Surface and Secreted Proteins and Their Extracellular Domains. Journal of Molecular Biology 430, 3353–3411. 10.1016/j.jmb.2018.06.002. [DOI] [PubMed] [Google Scholar]
114.Zhan X-L, Clemens JC, Neves G, Hattori D, Flanagan JJ, Hummel T, Vasconcelos ML, Chess A, and Zipursky SL (2004). Analysis of Dscam diversity in regulating axon guidance in Drosophila mushroom bodies. Neuron 43, 673–686. 10.1016/j.neuron.2004.07.020. [DOI] [PubMed] [Google Scholar]
115.Hao Y, Stuart T, Kowalski MH, Choudhary S, Hoffman P, Hartman A, Srivastava A, Molla G, Madad S, Fernandez-Granda C, et al. (2024). Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat. Biotechnol 42, 293–304. 10.1038/s41587-023-01767-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
116.Stocker RF, Heimbeck G, Gendre N, and de Belle JS (1997). Neuroblast ablation in Drosophila P[GAL4] lines reveals origins of olfactory interneurons. J. Neurobiol 32, 443–456. . [DOI] [PubMed] [Google Scholar]
117.Kvon EZ, Kazmar T, Stampfel G, Yáñez-Cuna JO, Pagani M, Schernhuber K, Dickson BJ, and Stark A. (2014). Genome-scale functional characterization of Drosophila developmental enhancers in vivo. Nature 512, 91–95. 10.1038/nature13395. [DOI] [PubMed] [Google Scholar]
118.Jenett A, Rubin GM, Ngo T-TB, Shepherd D, Murphy C, Dionne H, Pfeiffer BD, Cavallaro A, Hall D, Jeter J, et al. (2012). A GAL4-driver line resource for Drosophila neurobiology. Cell Rep. 2, 991–1001. 10.1016/j.celrep.2012.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
119.Connolly JB, Roberts IJ, Armstrong JD, Kaiser K, Forte M, Tully T, and O’Kane CJ (1996). Associative learning disrupted by impaired Gs signaling in Drosophila mushroom bodies. Science 274, 2104–2107. 10.1126/science.274.5295.2104. [DOI] [PubMed] [Google Scholar]
120.Pfeiffer BD, Truman JW, and Rubin GM (2012). Using translational enhancers to increase transgene expression in Drosophila. Proc. Natl. Acad. Sci. U. S. A 109, 6626–6631. 10.1073/pnas.1204520109. [DOI] [PMC free article] [PubMed] [Google Scholar]
121.Lee T, and Luo L. (1999). Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22, 451–461. 10.1016/s0896-6273(00)80701-1. [DOI] [PubMed] [Google Scholar]
122.Han DD, Stein D, and Stevens LM (2000). Investigating the function of follicular subpopulations during Drosophila oogenesis through hormone-dependent enhancer-targeted cell ablation. Development 127, 573–583. 10.1242/dev.127.3.573. [DOI] [PubMed] [Google Scholar]
123.Zirin J, Hu Y, Liu L, Yang-Zhou D, Colbeth R, Yan D, Ewen-Campen B, Tao R, Vogt E, VanNest S, et al. (2020). Large-scale transgenic Drosophila resource collections for loss- and gain-of-function studies. Genetics 214, 755–767. 10.1534/genetics.119.302964. [DOI] [PMC free article] [PubMed] [Google Scholar]
124.Bischof J, Björklund M, Furger E, Schertel C, Taipale J, and Basler K. (2013). A versatile platform for creating a comprehensive UAS-ORFeome library in Drosophila. Development 140, 2434–2442. 10.1242/dev.088757. [DOI] [PubMed] [Google Scholar]
125.Cachero S, Gkantia M, Bates AS, Frechter S, Blackie L, McCarthy A, Sutcliffe B, Strano A, Aso Y, and Jefferis GSXE (2020). BAcTrace, a tool for retrograde tracing of neuronal circuits in Drosophila. Nat. Methods 17, 1254–1261. 10.1038/s41592-020-00989-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
126.Tirian L, and Dickson BJ (2017). The VT GAL4, LexA, and split-GAL4 driver line collections for targeted expression in the Drosophila nervous system. medRxiv, 198648. 10.1101/198648. [DOI] [Google Scholar]
127.Xie Q, Wu B, Li J, Xu C, Li H, Luginbuhl DJ, Wang X, Ward A, and Luo L. (2019). Transsynaptic Fish-lips signaling prevents misconnections between nonsynaptic partner olfactory neurons. Proc. Natl. Acad. Sci. U. S. A 116, 16068–16073. 10.1073/pnas.1905832116. [DOI] [PMC free article] [PubMed] [Google Scholar]
128.Pitman JL, Huetteroth W, Burke CJ, Krashes MJ, Lai S-L, Lee T, and Waddell S. (2011). A pair of inhibitory neurons are required to sustain labile memory in the Drosophila mushroom body. Curr. Biol 21, 855–861. 10.1016/j.cub.2011.03.069. [DOI] [PMC free article] [PubMed] [Google Scholar]
129.Lai S-L, and Lee T. (2006). Genetic mosaic with dual binary transcriptional systems in Drosophila. Nat. Neurosci 9, 703–709. 10.1038/nn1681. [DOI] [PubMed] [Google Scholar]
130.Potter CJ, Tasic B, Russler EV, Liang L, and Luo L. (2010). The Q system: a repressible binary system for transgene expression, lineage tracing, and mosaic analysis. Cell 141, 536–548. 10.1016/j.cell.2010.02.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
131.Hong W, Mosca TJ, and Luo L. (2012). Teneurins instruct synaptic partner matching in an olfactory map. Nature 484, 201–207. 10.1038/nature10926. [DOI] [PMC free article] [PubMed] [Google Scholar]
132.Markstein M, Pitsouli C, Villalta C, Celniker SE, and Perrimon N. (2008). Exploiting position effects and the gypsy retrovirus insulator to engineer precisely expressed transgenes. Nat. Genet 40, 476–483. 10.1038/ng.101. [DOI] [PMC free article] [PubMed] [Google Scholar]
133.Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, Schmid B, et al. (2012). Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682. 10.1038/nmeth.2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
134.R Foundation for Statistical Computing (2023). R: A language and Environment for Statistical Computing.
135.Plaza SM, Clements J, Dolafi T, Umayam L, Neubarth NN, Scheffer LK, and Berg S. (2022). neuPrint: An open access tool for EM connectomics. Front. Neuroinform 16, 896292. 10.3389/fninf.2022.896292. [DOI] [PMC free article] [PubMed] [Google Scholar]
136.Bates AS, Manton JD, Jagannathan SR, Costa M, Schlegel P, Rohlfing T, and Jefferis GS (2020). The natverse, a versatile toolbox for combining and analysing neuroanatomical data. Elife 9. 10.7554/eLife.53350. [DOI] [PMC free article] [PubMed] [Google Scholar]
137.Hao Y, Hao S, Andersen-Nissen E, Mauck WM 3rd, Zheng S, Butler A, Lee MJ, Wilk AJ, Darby C, Zager M, et al. (2021). Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e29. 10.1016/j.cell.2021.04.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
138.Satija R, Farrell JA, Gennert D, Schier AF, and Regev A. (2015). Spatial reconstruction of single-cell gene expression data. Nat. Biotechnol 33, 495–502. 10.1038/nbt.3192. [DOI] [PMC free article] [PubMed] [Google Scholar]
139.Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM 3rd, Hao Y, Stoeckius M, Smibert P, and Satija R. (2019). Comprehensive integration of single-cell data. Cell 177, 1888–1902.e21. 10.1016/j.cell.2019.05.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
140.Neuwirth E. (2022). RColorBrewer: ColorBrewer Palettes. [Google Scholar]
141.Wolf FA, Angerer P, and Theis FJ (2018). SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 15. 10.1186/s13059-017-1382-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
142.Gayoso A, Lopez R, Xing G, Boyeau P, Valiollah Pour Amiri V, Hong J, Wu K, Jayasuriya M, Mehlman E, Langevin M, et al. (2022). A Python library for probabilistic analysis of single-cell omics data. Nat. Biotechnol 40, 163–166. 10.1038/s41587-021-01206-w. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information

NIHMS2125319-supplement-Supplementary_Information.pdf^{(2.3MB, pdf)}

[R1] 1.Marr D. (1969). A theory of cerebellar cortex. J. Physiol 202, 437–470. 10.1113/jphysiol.1969.sp008820. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] 2.Albus JS (1971). A theory of cerebellar function. Math. Biosci 10, 25–61. 10.1016/0025-5564(71)90051-4. [DOI] [Google Scholar]

[R3] 3.Ito M. (1970). Neurophysiological aspects of the cerebellar motor control system. Int. J. Neurol 7, 162–176. [PubMed] [Google Scholar]

[R4] 4.Chabrol FP, Arenz A, Wiechert MT, Margrie TW, and DiGregorio DA (2015). Synaptic diversity enables temporal coding of coincident multisensory inputs in single neurons. Nat. Neurosci 18, 718–727. 10.1038/nn.3974. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] 5.Huang C-C, Sugino K, Shima Y, Guo C, Bai S, Mensh BD, Nelson SB, and Hantman AW (2013). Convergence of pontine and proprioceptive streams onto multimodal cerebellar granule cells. Elife 2, e00400. 10.7554/eLife.00400. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Gruntman E, and Turner GC (2013). Integration of the olfactory code across dendritic claws of single mushroom body neurons. Nat. Neurosci 16, 1821–1829. 10.1038/nn.3547. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] 7.Ishikawa T, Shimuta M, and Häusser M. (2015). Multimodal sensory integration in single cerebellar granule cells in vivo. Elife 4. 10.7554/eLife.12916. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Cayco-Gajic NA, Clopath C, and Silver RA (2017). Sparse synaptic connectivity is required for decorrelation and pattern separation in feedforward networks. Nat. Commun 8, 1116. 10.1038/s41467-017-01109-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Litwin-Kumar A, Harris KD, Axel R, Sompolinsky H, and Abbott LF (2017). Optimal Degrees of Synaptic Connectivity. Neuron 93, 1153–1164.e7. 10.1016/j.neuron.2017.01.030. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Cayco-Gajic NA, and Silver RA (2019). Re-evaluating Circuit Mechanisms Underlying Pattern Separation. Neuron 101, 584–602. 10.1016/j.neuron.2019.01.044. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] 11.Turner GC, Bazhenov M, and Laurent G. (2008). Olfactory representations by Drosophila mushroom body neurons. J. Neurophysiol 99, 734–746. 10.1152/jn.01283.2007. [DOI] [PubMed] [Google Scholar]

[R12] 12.Farris SM (2011). Are mushroom bodies cerebellum-like structures? Arthropod Struct. Dev 40, 368–379. 10.1016/j.asd.2011.02.004. [DOI] [PubMed] [Google Scholar]

[R13] 13.Sawtell NB (2010). Multimodal integration in granule cells as a basis for associative plasticity and sensory prediction in a cerebellum-like circuit. Neuron 66, 573–584. 10.1016/j.neuron.2010.04.018. [DOI] [PubMed] [Google Scholar]

[R14] 14.Hobbs MJ, and Young JZ (1973). A cephalopod cerebellum. Brain Res. 55, 424–430. 10.1016/0006-8993(73)90307-7. [DOI] [PubMed] [Google Scholar]

[R15] 15.Leutgeb JK, Leutgeb S, Moser M-B, and Moser EI (2007). Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science 315, 961–966. 10.1126/science.1135801. [DOI] [PubMed] [Google Scholar]

[R16] 16.Srinivasan S, and Stevens CF (2018). The distributed circuit within the piriform cortex makes odor discrimination robust. J. Comp. Neurol 526, 2725–2743. 10.1002/cne.24492. [DOI] [PubMed] [Google Scholar]

[R17] 17.Aso Y, Grübel K, Busch S, Friedrich AB, Siwanowicz I, and Tanimoto H. (2009). The mushroom body of adult Drosophila characterized by GAL4 drivers. J. Neurogenet 23, 156–172. 10.1080/01677060802471718. [DOI] [PubMed] [Google Scholar]

[R18] 18.Caron SJC, Ruta V, Abbott LF, and Axel R. (2013). Random convergence of olfactory inputs in the Drosophila mushroom body. Nature 497, 113–117. 10.1038/nature12063. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Jefferis GS, Marin EC, Stocker RF, and Luo L. (2001). Target neuron prespecification in the olfactory map of Drosophila. Nature 414, 204–208. 10.1038/35102574. [DOI] [PubMed] [Google Scholar]

[R20] 20.Ahmed M, Rajagopalan AE, Pan Y, Li Y, Williams DL, Pedersen EA, Thakral M, Previero A, Close KC, Christoforou CP, et al. (2023). Input density tunes Kenyon cell sensory responses in the Drosophila mushroom body. Curr. Biol 33, 2742–2760.e12. 10.1016/j.cub.2023.05.064. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] 21.Elkahlah NA, Rogow JA, Ahmed M, and Clowney EJ (2020). Presynaptic developmental plasticity allows robust sparse wiring of the drosophila mushroom body. Elife 9. 10.7554/eLife.52278. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.Eichler K, Li F, Litwin-Kumar A, Park Y, Andrade I, Schneider-Mizell CM, Saumweber T, Huser A, Eschbach C, Gerber B, et al. (2017). The complete connectome of a learning and memory centre in an insect brain. Nature 548, 175–182. 10.1038/nature23455. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] 23.Li F, Lindsey JW, Marin EC, Otto N, Dreher M, Dempsey G, Stark I, Bates AS, Pleijzier MW, Schlegel P, et al. (2020). The connectome of the adult Drosophila mushroom body provides insights into function. Elife 9. 10.7554/eLife.62576. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.Murthy M, Fiete I, and Laurent G. (2008). Testing odor response stereotypy in the Drosophila mushroom body. Neuron 59, 1009–1023. 10.1016/j.neuron.2008.07.040. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R25] 25.Zheng Z, Li F, Fisher C, Ali IJ, Sharifi N, Calle-Schuler S, Hsu J, Masoodpanah N, Kmecova L, Kazimiers T, et al. (2022). Structured sampling of olfactory input by the fly mushroom body. Curr. Biol 32, 3334–3349.e6. 10.1016/j.cub.2022.06.031. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Ito K, Awano W, Suzuki K, Hiromi Y, and Yamamoto D. (1997). The Drosophila mushroom body is a quadruple structure of clonal units each of which contains a virtually identical set of neurones and glial cells. Development 124, 761–771. 10.1242/dev.124.4.761. [DOI] [PubMed] [Google Scholar]

[R27] 27.Kunz T, Kraft KF, Technau GM, and Urbach R. (2012). Origin of Drosophila mushroom body neuroblasts and generation of divergent embryonic lineages. Development 139, 2510–2522. 10.1242/dev.077883. [DOI] [PubMed] [Google Scholar]

[R28] 28.Lee T, Lee A, and Luo L. (1999). Development of the Drosophila mushroom bodies: Sequential generation of three distinct types of neurons from a neuroblast. Development 126, 4065–4076. [DOI] [PubMed] [Google Scholar]

[R29] 29.Yu H-H, Kao C-F, He Y, Ding P, Kao J-C, and Lee T. (2010). A complete developmental sequence of a Drosophila neuronal lineage as revealed by twin-spot MARCM. PLoS Biol. 8, e1000461. 10.1371/journal.pbio.1000461. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R30] 30.Lin S, Kao C-F, Yu H-H, Huang Y, and Lee T. (2012). Lineage analysis of Drosophila lateral antennal lobe neurons reveals notch-dependent binary temporal fate decisions. PLoS Biol. 10, e1001425. 10.1371/journal.pbio.1001425. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Ganguly I, Heckman EL, Litwin-Kumar A, Clowney EJ, and Behnia R. (2024). Diversity of visual inputs to Kenyon cells of the Drosophila mushroom body. Nat. Commun 15, 5698. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Wong AM, Wang JW, and Axel R. (2002). Spatial representation of the glomerular map in the Drosophila protocerebrum. Cell 109, 229–241. 10.1016/s0092-8674(02)00707-9. [DOI] [PubMed] [Google Scholar]

[R33] 33.Marin EC, Jefferis GSXE, Komiyama T, Zhu H, and Luo L. (2002). Representation of the glomerular olfactory map in the Drosophila brain. Cell 109, 243–255. 10.1016/S0092-8674(02)00700-6. [DOI] [PubMed] [Google Scholar]

[R34] 34.Tanaka NK, Awasaki T, Shimada T, and Ito K. (2004). Integration of chemosensory pathways in the Drosophila second-order olfactory centers. Curr. Biol 14, 449–457. 10.1016/j.cub.2004.03.006. [DOI] [PubMed] [Google Scholar]

[R35] 35.Jefferis GSXE, Vyas RM, Berdnik D, Ramaekers A, Stocker RF, Tanaka NK, Ito K, and Luo L. (2004). Developmental origin of wiring specificity in the olfactory system of Drosophila. Development 131, 117–130. 10.1242/dev.00896. [DOI] [PubMed] [Google Scholar]

[R36] 36.Yang J-Y, O’Connell TF, Hsu W-MM, Bauer MS, Dylla KV, Sharpee TO, and Hong EJ (2023). Restructuring of olfactory representations in the fly brain around odor relationships in natural sources. bioRxiv, 2023.02.15.528627. 10.1101/2023.02.15.528627. [DOI] [Google Scholar]

[R37] 37.Hayashi TT, MacKenzie AJ, Ganguly I, Ellis KE, Smihula HM, Jacob MS, Litwin-Kumar A, and Caron SJC (2022). Mushroom body input connections form independently of sensory activity in Drosophila melanogaster. Curr. Biol 0. 10.1016/j.cub.2022.07.055. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.Ellis KE, Bervoets S, Smihula H, Ganguly I, Vigato E, Auer TO, Benton R, Litwin-Kumar A, and Caron SJC (2024). Evolution of connectivity architecture in the Drosophila mushroom body. Nat. Commun 15, 4872. 10.1038/s41467-024-48839-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] 39.Lin H-H, Lai JS-Y, Chin A-L, Chen Y-C, and Chiang A-S (2007). A map of olfactory representation in the Drosophila mushroom body. Cell 128, 1205–1217. 10.1016/j.cell.2007.03.006. [DOI] [PubMed] [Google Scholar]

[R40] 40.Campbell RAA, Honegger KS, Qin H, Li W, Demir E, and Turner GC (2013). Imaging a population code for odor identity in the Drosophila mushroom body. J. Neurosci 33, 10568–10581. 10.1523/JNEUROSCI.0682-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R41] 41.Bengtsson F, and Jörntell H. (2009). Sensory transmission in cerebellar granule cells relies on similarly coded mossy fiber inputs. Proc. Natl. Acad. Sci. U. S. A 106, 2389–2394. 10.1073/pnas.0808428106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R42] 42.Shuster SA, Wagner MJ, Pan-Doh N, Ren J, Grutzner SM, Beier KT, Kim TH, Schnitzer MJ, and Luo L. (2021). The relationship between birth timing, circuit wiring, and physiological response properties of cerebellar granule cells. Proc. Natl. Acad. Sci. U. S. A 118. 10.1073/pnas.2101826118. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] 43.Kennedy A, Wayne G, Kaifosh P, Alviña K, Abbott LF, and Sawtell NB (2014). A temporal basis for predicting the sensory consequences of motor commands in an electric fish. Nat. Neurosci 17, 416–422. 10.1038/nn.3650. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] 44.Gilmer JI, and Person AL (2017). Morphological Constraints on Cerebellar Granule Cell Combinatorial Diversity. J. Neurosci 37, 12153–12166. 10.1523/JNEUROSCI.0588-17.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] 45.Nguyen TM, Thomas LA, Rhoades JL, Ricchi I, Yuan XC, Sheridan A, Hildebrand DGC, Funke J, Regehr WG, and Lee W-CA (2023). Structured cerebellar connectivity supports resilient pattern separation. Nature 613, 543–549. 10.1038/s41586-022-05471-w. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R46] 46.Scheffer LK, Xu CS, Januszewski M, Lu Z, Takemura S-Y, Hayworth KJ, Huang GB, Shinomiya K, Maitlin-Shepard J, Berg S, et al. (2020). A connectome and analysis of the adult Drosophila central brain. Elife 9. 10.7554/eLife.57443. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Zhu S, Chiang A-S, and Lee T. (2003). Development of the Drosophila mushroom bodies: elaboration, remodeling and spatial organization of dendrites in the calyx. Development 130, 2603–2610. 10.1242/dev.00466. [DOI] [PubMed] [Google Scholar]

[R48] 48.Jefferis GSXE, Potter CJ, Chan AM, Marin EC, Rohlfing T, Maurer CR Jr, and Luo L. (2007). Comprehensive maps of Drosophila higher olfactory centers: spatially segregated fruit and pheromone representation. Cell 128, 1187–1203. 10.1016/j.cell.2007.01.040. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Zheng Z, Lauritzen JS, Perlman E, Robinson CG, Nichols M, Milkie D, Torrens O, Price J, Fisher CB, Sharifi N, et al. (2018). A Complete Electron Microscopy Volume of the Brain of Adult Drosophila melanogaster. Cell 174, 730–743.e22. 10.1016/j.cell.2018.06.019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R50] 50.Li H, Horns F, Xie Q, Xie Q, Li T, Luginbuhl DJ, Luo L, and Quake SR (2017). Classifying Drosophila Olfactory Projection Neuron Subtypes by Single-Cell RNA Sequencing. Cell 171, 1206.e22–1220.e22. 10.1016/j.cell.2017.10.019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] 51.Xie Q, Brbic M, Horns F, Kolluru SS, Jones RC, Li J, Reddy AR, Xie A, Kohani S, Li Z, et al. (2021). Temporal evolution of single-cell transcriptomes of Drosophila olfactory projection neurons. Elife 10. 10.7554/eLife.63450. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] 52.Carrillo RA, Özkan E, Menon KP, Nagarkar-Jaiswal S, Lee PT, Jeon M, Birnbaum ME, Bellen HJ, Garcia KC, and Zinn K. (2015). Control of Synaptic Connectivity by a Network of Drosophila IgSF Cell Surface Proteins. Cell 163, 1770–1782. 10.1016/j.cell.2015.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R53] 53.Xu C, Theisen E, Maloney R, Peng J, Santiago I, Yapp C, Werkhoven Z, Rumbaut E, Shum B, Tarnogorska D, et al. (2019). Control of Synaptic Specificity by Establishing a Relative Preference for Synaptic Partners. Neuron 103, 865–877.e7. 10.1016/j.neuron.2019.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] 54.Yoo J, Dombrovski M, Mirshahidi P, Nern A, LoCascio SA, Zipursky SL, and Kurmangaliyev YZ (2023). Brain wiring determinants uncovered by integrating connectomes and transcriptomes. Curr. Biol 33, 3998–4005.e6. 10.1016/j.cub.2023.08.020. [DOI] [PubMed] [Google Scholar]

[R55] 55.Özkan E, Carrillo RA, Eastman CL, Weiszmann R, Waghray D, Johnson KG, Zinn K, Celniker SE, and Garcia KC (2013). An extracellular interactome of immunoglobulin and LRR proteins reveals receptor-ligand networks. Cell 154, 228. 10.1016/j.cell.2013.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] 56.Tan L, Zhang KX, Pecot MY, Nagarkar-Jaiswal S, Lee PT, Takemura SY, McEwen JM, Nern A, Xu S, Tadros W, et al. (2015). Ig Superfamily Ligand and Receptor Pairs Expressed in Synaptic Partners in Drosophila. Cell 163, 1756–1769. 10.1016/j.cell.2015.11.021.Ig. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] 57.Bates AS, Schlegel P, Roberts RJV, Drummond N, Tamimi IFM, Turnbull R, Zhao X, Marin EC, Popovici PD, Dhawan S, et al. (2020). Complete Connectomic Reconstruction of Olfactory Projection Neurons in the Fly Brain. Curr. Biol 30, 3183–3199.e6. 10.1016/j.cub.2020.06.042. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R58] 58.Yusuyama K, Meinertzhagen IA, and Schürmann FW (2002). Synaptic organization of the mushroom body calyx in Drosophila melanogaster. J. Comp. Neurol 445, 211–226. 10.1002/cne.10155. [DOI] [PubMed] [Google Scholar]

[R59] 59.Leiss F, Groh C, Butcher NJ, Meinertzhagen IA, and Tavosanis G. (2009). Synaptic organization in the adult Drosophila mushroom body calyx. J. Comp. Neurol 517, 808–824. 10.1002/cne.22184. [DOI] [PubMed] [Google Scholar]

[R60] 60.Yang K, Liu T, Wang Z, Liu J, Shen Y, Pan X, Wen R, Xie H, Ruan Z, Tan Z, et al. (2022). Classifying Drosophila olfactory projection neuron boutons by quantitative analysis of electron microscopic reconstruction. iScience 25, 104180. 10.1016/j.isci.2022.104180. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] 61.Butcher NJ, Friedrich AB, Lu Z, Tanimoto H, and Meinertzhagen IA (2012). Different classes of input and output neurons reveal new features in microglomeruli of the adult Drosophila mushroom body calyx. J. Comp. Neurol 520, 2185–2201. 10.1002/cne.23037. [DOI] [PubMed] [Google Scholar]

[R62] 62.Kurusu M, Cording A, Taniguchi M, Menon K, Suzuki E, and Zinn K. (2008). A screen of cell-surface molecules identifies leucine-rich repeat proteins as key mediators of synaptic target selection. Neuron 59, 972–985. 10.1016/j.neuron.2008.07.037. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R63] 63.Muthukumar AK, Stork T, and Freeman MR (2014). Activity-dependent regulation of astrocyte GAT levels during synaptogenesis. Nat. Neurosci 17, 1340–1350. 10.1038/nn.3791. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R64] 64.Croset V, Treiber CD, and Waddell S. (2018). Cellular diversity in the Drosophila midbrain revealed by single-cell transcriptomics. Elife 7, 1–31. 10.7554/eLife.34550. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R65] 65.Lee G, Foss M, Goodwin SF, Carlo T, Taylor BJ, and Hall JC (2000). Spatial, temporal, and sexually dimorphic expression patterns of the fruitless gene in the Drosophila central nervous system. Journal of Neurobiology 43, 404–426. . [DOI] [PubMed] [Google Scholar]

[R66] 66.Shih MFM, Davis FP, Henry GL, and Dubnau J. (2019). Nuclear transcriptomes of the seven neuronal cell types that constitute the drosophila mushroom bodies. G3: Genes, Genomes, Genetics 9, 81–94. 10.1534/g3.118.200726. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R67] 67.Vogt K, Aso Y, Hige T, Knapek S, Ichinose T, Friedrich AB, Turner GC, Rubin GM, and Tanimoto H. (2016). Direct neural pathways convey distinct visual information to drosophila mushroom bodies. Elife 5, 1–13. 10.7554/eLife.14009. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R68] 68.Li J, Mahoney BD, Jacob MS, and Caron SJC (2020). Visual Input into The Drosophila Melanogaster Mushroom Body. SSRN Electronic Journal, 1–47. 10.2139/ssrn.3595714. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R69] 69.Jain S, Lin Y, Kurmangaliyev YZ, Valdes-Aleman J, LoCascio SA, Mirshahidi P, Parrington B, and Zipursky SL (2022). A global timing mechanism regulates cell-type-specific wiring programmes. Nature 603, 112–118. 10.1038/s41586-022-04418-5. [DOI] [PubMed] [Google Scholar]

[R70] 70.Kurmangaliyev YZ, Yoo J, Valdes-Aleman J, Sanfilippo P, and Zipursky SL (2020). Transcriptional Programs of Circuit Assembly in the Drosophila Visual System. Neuron 108, 1045–1057.e6. 10.1016/j.neuron.2020.10.006. [DOI] [PubMed] [Google Scholar]

[R71] 71.Alyagor I, Berkun V, Keren-Shaul H, Marmor-Kollet N, David E, Mayseless O, Issman-Zecharya N, Amit I, and Schuldiner O. (2018). Combining developmental and perturbation-seq uncovers transcriptional modules orchestrating neuronal remodeling. Dev. Cell 47, 38–52.e6. 10.1016/j.devcel.2018.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R72] 72.Ahmed M. (2023). Generation of sparse, combinatorial wiring for sensory coding. 10.7302/8559. [DOI] [Google Scholar]

[R73] 73.Li H, Watson A, Olechwier A, Anaya M, Sorooshyari SK, Harnett DP, Lee H-KP, Vielmetter J, Fares MA, Garcia KC, et al. (2017). Deconstruction of the beaten Path-Sidestep interaction network provides insights into neuromuscular system development. Elife 6. 10.7554/eLife.28111. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R74] 74.Barish S, Nuss S, Strunilin I, Bao S, Mukherjee S, Jones CD, and Volkan PC (2018). Combinations of DIPs and Dprs control organization of olfactory receptor neuron terminals in Drosophila. PLOS Genetics 14, 1–33. 10.1371/journal.pgen.1007560. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R75] 75.Brovero SG, Fortier JC, Hu H, Lovejoy PC, Newell NR, Palmateer CM, Tzeng R-Y, Lee P-T, Zinn K, and Arbeitman MN (2021). Investigation of Drosophila fruitless neurons that express Dpr/DIP cell adhesion molecules. Elife 10. 10.7554/eLife.63101. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R76] 76.Sanes JR, and Zipursky SL (2020). Synaptic Specificity, Recognition Molecules, and Assembly of Neural Circuits. Cell 181, 536–556. 10.1016/j.cell.2020.04.008. [DOI] [PubMed] [Google Scholar]

[R77] 77.Hiesinger PR (2021). Brain wiring with composite instructions. Bioessays 43, e2000166. 10.1002/bies.202000166. [DOI] [PubMed] [Google Scholar]

[R78] 78.Hassan BA, and Hiesinger PR (2015). Beyond Molecular Codes: Simple Rules to Wire Complex Brains. Cell 163, 285–291. 10.1016/j.cell.2015.09.031. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R79] 79.Menon KP, Kulkarni V, Takemura S-Y, Anaya M, and Zinn K. (2019). Interactions between Dpr11 and DIP-γ control selection of amacrine neurons in Drosophila color vision circuits. Elife 8. 10.7554/eLife.48935. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R80] 80.Bornstein B, Meltzer H, Adler R, Alyagor I, Berkun V, Cummings G, Reh F, Keren-Shaul H, David E, Riemensperger T, et al. (2021). Transneuronal Dpr12/DIP-δ interactions facilitate compartmentalized dopaminergic innervation of Drosophila mushroom body axons. EMBO J. 40, e105763. 10.15252/embj.2020105763. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R81] 81.Courgeon M, and Desplan C. (2019). Coordination between stochastic and deterministic specification in the Drosophila visual system. Science 366, eaay6727. 10.1126/science.aay6727. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R82] 82.Cheng S, Park Y, Kurleto JD, Jeon M, Zinn K, Thornton JW, and Özkan E. (2019). Family of neural wiring receptors in bilaterians defined by phylogenetic, biochemical, and structural evidence. Proc. Natl. Acad. Sci. U. S. A 116, 9837–9842. 10.1073/pnas.1818631116. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R83] 83.Morano NC, Lopez DH, Meltzer H, Sergeeva AP, Katsamba PS, Rostam KD, Gupta HP, Becker JE, Bornstein B, Cosmanescu F, et al. (2025). Members of the DIP and Dpr adhesion protein families use cis inhibition to shape neural development in Drosophila. PLoS Biol. 23, e3003030. 10.1371/journal.pbio.3003030. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R84] 84.Andreatta M, and Carmona SJ (2021). UCell: Robust and scalable single-cell gene signature scoring. Comput. Struct. Biotechnol. J 19, 3796–3798. 10.1016/j.csbj.2021.06.043. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R85] 85.Palmateer CM, Artikis C, Brovero SG, Friedman B, Gresham A, and Arbeitman MN (2023). Single-cell transcriptome profiles of Drosophila fruitless-expressing neurons from both sexes. Elife 12. 10.7554/eLife.78511. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R86] 86.Özel MN, Simon F, Jafari S, Holguera I, Chen Y-C, Benhra N, El-Danaf RN, Kapuralin K, Malin JA, Konstantinides N, et al. (2021). Neuronal diversity and convergence in a visual system developmental atlas. Nature 589, 88–95. 10.1038/s41586-020-2879-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R87] 87.Yagi R, Mabuchi Y, Mizunami M, and Tanaka NK (2016). Convergence of multimodal sensory pathways to the mushroom body calyx in Drosophila melanogaster. Sci. Rep 6, 1–8. 10.1038/srep29481. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R88] 88.Rosenblatt F. (1958). THE PERCEPTRON: A PROBABILISTIC MODEL FOR INFORMATION STORAGE AND ORGANIZATION IN THE BRAIN. Psychol. Rev 65, 386–408. [DOI] [PubMed] [Google Scholar]

[R89] 89.De Belle JS, and Heisenberg M. (1994). Associative odor learning in Drosophila abolished by chemical ablation of mushroom bodies. Science 263, 692–695. 10.1126/science.8303280. [DOI] [PubMed] [Google Scholar]

[R90] 90.Masse NY, Turner GC, and Jefferis GSXE (2009). Olfactory Information Processing in Drosophila. Curr. Biol 19, R700–R713. 10.1016/j.cub.2009.06.026. [DOI] [PubMed] [Google Scholar]

[R91] 91.Laurent G, and Naraghi M. (1994). Odorant-induced oscillations in the mushroom bodies of the locust. J. Neurosci 14, 2993–3004. 10.1523/jneurosci.14-05-02993.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R92] 92.Di Bella DJ, Habibi E, Stickels RR, Scalia G, Brown J, Yadollahpour P, Yang SM, Abbate C, Biancalani T, Macosko EZ, et al. (2021). Molecular logic of cellular diversification in the mouse cerebral cortex. Nature 595, 554–559. 10.1038/s41586-021-03670-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R93] 93.Chen Y, Chen X, Baserdem B, Zhan H, Li Y, Davis MB, Kebschull JM, Zador AM, Koulakov AA, and Albeanu DF (2022). High-throughput sequencing of single neuron projections reveals spatial organization in the olfactory cortex. Cell 185, 4117–4134.e28. 10.1016/j.cell.2022.09.038. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R94] 94.Yu Y-C, Bultje RS, Wang X, and Shi S-H (2009). Specific synapses develop preferentially among sister excitatory neurons in the neocortex. Nature 458, 501–504. 10.1038/nature07722. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R95] 95.Li Y, Lu H, Cheng P-L, Ge S, Xu H, Shi S-H, and Dan Y. (2012). Clonally related visual cortical neurons show similar stimulus feature selectivity. Nature 486, 118–121. 10.1038/nature11110. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R96] 96.Kasthuri N, Hayworth KJ, Berger DR, Schalek RL, Conchello JA, Knowles-Barley S, Lee D, Vázquez-Reina A, Kaynig V, Jones TR, et al. (2015). Saturated Reconstruction of a Volume of Neocortex. Cell 162, 648–661. 10.1016/j.cell.2015.06.054. [DOI] [PubMed] [Google Scholar]

[R97] 97.Wolterhoff N, and Hiesinger PR (2024). Synaptic promiscuity in brain development. Curr. Biol 34, R102–R116. 10.1016/j.cub.2023.12.037. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R98] 98.Bekkers JM, and Stevens CF (1991). Excitatory and inhibitory autaptic currents in isolated hippocampal neurons maintained in cell culture. Proc. Natl. Acad. Sci. U. S. A 88, 7834–7838. 10.1073/pnas.88.17.7834. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R99] 99.Van der Loos H, and Glaser EM (1972). Autapses in neocortex cerebri: synapses between a pyramidal cell’s axon and its own dendrites. Brain Res. 48, 355–360. 10.1016/0006-8993(72)90189-8. [DOI] [PubMed] [Google Scholar]

[R100] 100.Grueber WB, and Sagasti A. (2010). Self-avoidance and tiling: Mechanisms of dendrite and axon spacing. Cold Spring Harb. Perspect. Biol 2, a001750. 10.1101/cshperspect.a001750. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R101] 101.Williams DL, Sikora VM, Hammer MA, Amin S, Brinjikji T, Brumley EK, Burrows CJ, Carrillo PM, Cromer K, Edwards SJ, et al. (2021). May the odds be ever in your favor: Non-deterministic mechanisms diversifying cell surface molecule expression. Front. Cell Dev. Biol 9, 720798. 10.3389/fcell.2021.720798. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R102] 102.Valdes-Aleman J, Fetter RD, Sales EC, Heckman EL, Venkatasubramanian L, Doe CQ, Landgraf M, Cardona A, and Zlatic M. (2021). Comparative Connectomics Reveals How Partner Identity, Location, and Activity Specify Synaptic Connectivity in Drosophila. Neuron 109, 105–122.e7. 10.1016/j.neuron.2020.10.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R103] 103.Kiral FR, Linneweber GA, Mathejczyk T, Georgiev SV, Wernet MF, Hassan BA, von Kleist M, and Hiesinger PR (2020). Autophagy-dependent filopodial kinetics restrict synaptic partner choice during Drosophila brain wiring. Nat. Commun 11, 1325. 10.1038/s41467-020-14781-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R104] 104.Kiral FR, Dutta SB, Linneweber GA, Hilgert S, Poppa C, Duch C, von Kleist M, Hassan BA, and Hiesinger PR (2021). Brain connectivity inversely scales with developmental temperature in Drosophila. Cell Rep. 37, 110145. 10.1016/j.celrep.2021.110145. [DOI] [PubMed] [Google Scholar]

[R105] 105.Osaka J, Ishii A, Wang X, Iwanaga R, Kawamura H, Akino S, Sugie A, Hakeda-Suzuki S, and Suzuki T. (2024). Complex formation of immunoglobulin superfamily molecules Side-IV and Beat-IIb regulates synaptic specificity. Cell Rep. 43, 113798. 10.1016/j.celrep.2024.113798. [DOI] [PubMed] [Google Scholar]

[R106] 106.Drotos AC, and Roberts MT (2024). Identifying neuron types and circuit mechanisms in the auditory midbrain. Hear. Res 442, 108938. 10.1016/j.heares.2023.108938. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R107] 107.Minsky M. (1952). A Neural-Analogue Calculator Based upon a Probability Model of Reinforcement. Preprint. [Google Scholar]

[R108] 108.Minsky M. (2016). Building my randomly wired neural network machine. [Google Scholar]

[R109] 109.Edelman GM (1989). The remembered present: A biological theory of consciousness. The remembered present: A biological theory of consciousness. [Google Scholar]

[R110] 110.Brovkina MV, Duffié R, Burtis AEC, and Clowney EJ (2021). Fruitless decommissions regulatory elements to implement cell-type-specific neuronal masculinization. PLoS Genet. 17, e1009338. 10.1371/journal.pgen.1009338. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R111] 111.Aso Y, Hattori D, Yu Y, Johnston RM, Iyer NA, Ngo TTB, Dionne H, Abbott LF, Axel R, Tanimoto H, et al. (2014). The neuronal architecture of the mushroom body provides a logic for associative learning. Elife 3, e04577. 10.7554/eLife.04577. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R112] 112.Zappia L, and Oshlack A. (2018). Clustering trees: a visualization for evaluating clusterings at multiple resolutions. GigaScience 7, giy083. 10.1093/gigascience/giy083. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R113] 113.Pei J, Kinch LN, and Grishin NV (2018). FlyXCDB—A Resource for Drosophila Cell Surface and Secreted Proteins and Their Extracellular Domains. Journal of Molecular Biology 430, 3353–3411. 10.1016/j.jmb.2018.06.002. [DOI] [PubMed] [Google Scholar]

[R114] 114.Zhan X-L, Clemens JC, Neves G, Hattori D, Flanagan JJ, Hummel T, Vasconcelos ML, Chess A, and Zipursky SL (2004). Analysis of Dscam diversity in regulating axon guidance in Drosophila mushroom bodies. Neuron 43, 673–686. 10.1016/j.neuron.2004.07.020. [DOI] [PubMed] [Google Scholar]

[R115] 115.Hao Y, Stuart T, Kowalski MH, Choudhary S, Hoffman P, Hartman A, Srivastava A, Molla G, Madad S, Fernandez-Granda C, et al. (2024). Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat. Biotechnol 42, 293–304. 10.1038/s41587-023-01767-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R116] 116.Stocker RF, Heimbeck G, Gendre N, and de Belle JS (1997). Neuroblast ablation in Drosophila P[GAL4] lines reveals origins of olfactory interneurons. J. Neurobiol 32, 443–456. . [DOI] [PubMed] [Google Scholar]

[R117] 117.Kvon EZ, Kazmar T, Stampfel G, Yáñez-Cuna JO, Pagani M, Schernhuber K, Dickson BJ, and Stark A. (2014). Genome-scale functional characterization of Drosophila developmental enhancers in vivo. Nature 512, 91–95. 10.1038/nature13395. [DOI] [PubMed] [Google Scholar]

[R118] 118.Jenett A, Rubin GM, Ngo T-TB, Shepherd D, Murphy C, Dionne H, Pfeiffer BD, Cavallaro A, Hall D, Jeter J, et al. (2012). A GAL4-driver line resource for Drosophila neurobiology. Cell Rep. 2, 991–1001. 10.1016/j.celrep.2012.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R119] 119.Connolly JB, Roberts IJ, Armstrong JD, Kaiser K, Forte M, Tully T, and O’Kane CJ (1996). Associative learning disrupted by impaired Gs signaling in Drosophila mushroom bodies. Science 274, 2104–2107. 10.1126/science.274.5295.2104. [DOI] [PubMed] [Google Scholar]

[R120] 120.Pfeiffer BD, Truman JW, and Rubin GM (2012). Using translational enhancers to increase transgene expression in Drosophila. Proc. Natl. Acad. Sci. U. S. A 109, 6626–6631. 10.1073/pnas.1204520109. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R121] 121.Lee T, and Luo L. (1999). Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22, 451–461. 10.1016/s0896-6273(00)80701-1. [DOI] [PubMed] [Google Scholar]

[R122] 122.Han DD, Stein D, and Stevens LM (2000). Investigating the function of follicular subpopulations during Drosophila oogenesis through hormone-dependent enhancer-targeted cell ablation. Development 127, 573–583. 10.1242/dev.127.3.573. [DOI] [PubMed] [Google Scholar]

[R123] 123.Zirin J, Hu Y, Liu L, Yang-Zhou D, Colbeth R, Yan D, Ewen-Campen B, Tao R, Vogt E, VanNest S, et al. (2020). Large-scale transgenic Drosophila resource collections for loss- and gain-of-function studies. Genetics 214, 755–767. 10.1534/genetics.119.302964. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R124] 124.Bischof J, Björklund M, Furger E, Schertel C, Taipale J, and Basler K. (2013). A versatile platform for creating a comprehensive UAS-ORFeome library in Drosophila. Development 140, 2434–2442. 10.1242/dev.088757. [DOI] [PubMed] [Google Scholar]

[R125] 125.Cachero S, Gkantia M, Bates AS, Frechter S, Blackie L, McCarthy A, Sutcliffe B, Strano A, Aso Y, and Jefferis GSXE (2020). BAcTrace, a tool for retrograde tracing of neuronal circuits in Drosophila. Nat. Methods 17, 1254–1261. 10.1038/s41592-020-00989-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R126] 126.Tirian L, and Dickson BJ (2017). The VT GAL4, LexA, and split-GAL4 driver line collections for targeted expression in the Drosophila nervous system. medRxiv, 198648. 10.1101/198648. [DOI] [Google Scholar]

[R127] 127.Xie Q, Wu B, Li J, Xu C, Li H, Luginbuhl DJ, Wang X, Ward A, and Luo L. (2019). Transsynaptic Fish-lips signaling prevents misconnections between nonsynaptic partner olfactory neurons. Proc. Natl. Acad. Sci. U. S. A 116, 16068–16073. 10.1073/pnas.1905832116. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R128] 128.Pitman JL, Huetteroth W, Burke CJ, Krashes MJ, Lai S-L, Lee T, and Waddell S. (2011). A pair of inhibitory neurons are required to sustain labile memory in the Drosophila mushroom body. Curr. Biol 21, 855–861. 10.1016/j.cub.2011.03.069. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R129] 129.Lai S-L, and Lee T. (2006). Genetic mosaic with dual binary transcriptional systems in Drosophila. Nat. Neurosci 9, 703–709. 10.1038/nn1681. [DOI] [PubMed] [Google Scholar]

[R130] 130.Potter CJ, Tasic B, Russler EV, Liang L, and Luo L. (2010). The Q system: a repressible binary system for transgene expression, lineage tracing, and mosaic analysis. Cell 141, 536–548. 10.1016/j.cell.2010.02.025. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R131] 131.Hong W, Mosca TJ, and Luo L. (2012). Teneurins instruct synaptic partner matching in an olfactory map. Nature 484, 201–207. 10.1038/nature10926. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R132] 132.Markstein M, Pitsouli C, Villalta C, Celniker SE, and Perrimon N. (2008). Exploiting position effects and the gypsy retrovirus insulator to engineer precisely expressed transgenes. Nat. Genet 40, 476–483. 10.1038/ng.101. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R133] 133.Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, Schmid B, et al. (2012). Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682. 10.1038/nmeth.2019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R134] 134.R Foundation for Statistical Computing (2023). R: A language and Environment for Statistical Computing.

[R135] 135.Plaza SM, Clements J, Dolafi T, Umayam L, Neubarth NN, Scheffer LK, and Berg S. (2022). neuPrint: An open access tool for EM connectomics. Front. Neuroinform 16, 896292. 10.3389/fninf.2022.896292. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R136] 136.Bates AS, Manton JD, Jagannathan SR, Costa M, Schlegel P, Rohlfing T, and Jefferis GS (2020). The natverse, a versatile toolbox for combining and analysing neuroanatomical data. Elife 9. 10.7554/eLife.53350. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R137] 137.Hao Y, Hao S, Andersen-Nissen E, Mauck WM 3rd, Zheng S, Butler A, Lee MJ, Wilk AJ, Darby C, Zager M, et al. (2021). Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e29. 10.1016/j.cell.2021.04.048. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R138] 138.Satija R, Farrell JA, Gennert D, Schier AF, and Regev A. (2015). Spatial reconstruction of single-cell gene expression data. Nat. Biotechnol 33, 495–502. 10.1038/nbt.3192. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R139] 139.Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM 3rd, Hao Y, Stoeckius M, Smibert P, and Satija R. (2019). Comprehensive integration of single-cell data. Cell 177, 1888–1902.e21. 10.1016/j.cell.2019.05.031. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R140] 140.Neuwirth E. (2022). RColorBrewer: ColorBrewer Palettes. [Google Scholar]

[R141] 141.Wolf FA, Angerer P, and Theis FJ (2018). SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 15. 10.1186/s13059-017-1382-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R142] 142.Gayoso A, Lopez R, Xing G, Boyeau P, Valiollah Pour Amiri V, Hong J, Wu K, Jayasuriya M, Mehlman E, Langevin M, et al. (2022). A Python library for probabilistic analysis of single-cell omics data. Nat. Biotechnol 40, 163–166. 10.1038/s41587-021-01206-w. [DOI] [PubMed] [Google Scholar]

PERMALINK

Spatial constraints and cell surface molecule depletion structure a randomly connected learning circuit

Emma M Thornton-Kolbe

Maria Ahmed

Finley R Gordon

Bogdan Sieriebriennikov

Donnell L Williams

Yerbol Z Kurmangaliyev

E Josephine Clowney

Summary

Introduction

Results

Projection neuron anatomy in the mushroom body calyx is orderly, biased, and predictable

Figure 1: Olfactory projection neurons have predictable anatomy in the calyx.

Individual Kenyon cells innervate restricted regions of the calyx

Figure 2: Kenyon cell dendrite position reflects neurogenic history and predicts repertoire of presynaptic inputs.

Within a Kenyon cell’s restricted region of the calyx, its inputs are randomized

Figure 3: Within developmentally defined spatial domains, Kenyon cell inputs are random.

Kenyon cells express a reduced set of cell surface molecules

Figure 4: Developmental analysis of transcription in Kenyon cell subtypes.

Analysis of transcriptional consistency within canonical Kenyon cell classes

Immunoglobulin superfamily genes are depleted in individual Kenyon cells

Figure 5: Expression of Ig-SF genes is broadly depleted in Kenyon cells.

When deprived of typical partners, Kenyon cells still synapse

Figure 6: Calyx development is robust to radical reduction of olfactory projection neuron diversity.

Ectopic Ig-SF expression attracts Kenyon cell claws onto boutons of projection neurons expressing cognate molecules

Figure 7: Ecoptic Ig-SF expression in Kenyon cells biases projection neuron contacts.

Discussion

STAR Methods

Experimental Model and Study Participant Details

Fly husbandry

Method Details

EM analysis

Downloading synapse locations from neuprint+

Example query:

Downloading, and trimming neuron skeletons

Cell type and developmental feature annotation

Subcellular feature annotation

Rotating coordinate space to match up with relevant calyx axes

Generation of random models of KC:PN connectivity

Conditional Input analysis

Bulk RNA sequencing

Flow cytometry

RNA isolation, library preparation, and sequencing for low-input bulk RNAseq on FAC-sorted neurons

Single-cell RNA-sequencing

Flow cytometry for single-cell RNA-sequencing of developing Kenyon cells

Labeling of adult Kenyon cells and preparation for scRNAseq

ScRNAseq prep, quality control metrics and clustering analysis of developing Kenyon cells

Selection of gene sets

Cell Surface and Secreted Molecules (CSS’s)

Ig-SF and LRR specificity molecules

GPCRs

Transcription factors

MiMIC reporter validation of sequencing data

Brain dissection, immunostaining, and confocal imaging

Dye filling Kenyon cells

Quantification and Statistical Analysis

Analysis of bulk RNAseq datasets

Analysis of single cell RNA sequencing

Analysis of residual transcriptional variation among Kenyon cells of the same type and stage

Transcriptional analysis of cell surface and secreted molecules in Kenyon cells

Expression of Ig-SF and LRR molecules

Re-analysis of published single cell RNA sequencing data

olfactory projection neurons

fruitless neurons

Image analysis

Analysis considerations

Area of the largest slice of the calyx

Projection neuron bouton counts

Projection neuron cell counts

Normalized Bruchpilot and Discs large intensity

Normalized Kenyon cell fluorescence in bouton ROIs

Kenyon cell claw counts

Statistical considerations

Supplementary Material

Acknowledgements

Footnotes

References

Associated Data

Supplementary Materials