The genomic basis of temporal niche evolution in a diurnal rodent

Summary

Patterns of diel activity – how animals allocate their activity throughout the 24hr daily cycle – play key roles in shaping the internal physiology of an animal and its relationship with the external environment^1–5. Although shifts in diel activity patterns have occurred numerous times over the course of amniote evolution⁶, the genomic correlates of such transitions remain unknown. Here, we use the African striped mouse (Rhabdomys pumilio), a species that transitioned from the ancestrally-nocturnal diel niche of its close relatives to a diurnal one^7–11, to define patterns of naturally-occurring molecular variation in diel niche traits. First, to facilitate genomic analyses, we generated a chromosome-level genome assembly of the striped mouse. Next, using transcriptomics, we show that the switch to daytime activity in this species is associated with a realignment of daily rhythms in peripheral tissues with respect to the light:dark cycle and the central circadian clock. To uncover selection pressures associated with this temporal niche shift, we perform comparative genomic analyses of closely related rodent species and find evidence of relaxation of purifying selection on striped mouse genes in the rod phototransduction pathway. In agreement with this, electroretinogram measurements demonstrate that striped mice have functional differences in dim light visual responses compared to nocturnal rodents. Taken together, our results show that striped mice have undergone a drastic change in circadian organisation and provide evidence that the visual system has been a major target of selection as this species transitioned to a novel temporal niche.

eTOC Blurb

Richardson et al. study R. pumilio, a rodent that transitioned from the ancestrally-nocturnal diel niche of its close relatives to a diurnal one. Using whole-genome sequencing, transcriptomics, comparative genomics, and retinal measurements of light response, they define patterns of naturally occurring molecular variation in diel niche traits.

Results and Discussion

The set of ecological interactions that define a species’ niche are distributed both spatially across its habitat as well as temporally across seasonal and diel (day-night) cycles. Although it is common for animals to show some flexibility in daily patterns of activity, many species exhibit a preferred diel temporal niche. A species’ diel niche profoundly shapes its ability to acquire information through its senses, its access to resources and concomitant foraging strategies, energy and water consumption, methods of predator avoidance and its means for both intraspecific and interspecific communication¹. Given this, diel niche is often reflected by a suite of morphological, physiological, and behavioural adaptations that are well-documented among diverse lineages^{2–5,12–16}. Critically, diel niche preference is fundamentally linked to organismal biology through internal circadian clocks, which establish 24hr rhythms in physiology and behaviour. Implementation and maintenance of diel niche therefore requires correct phasing of these rhythms with respect to the light:dark cycle.

While the molecular basis of some diel temporal niche traits has been studied in traditional model organisms like the laboratory mouse^17–19, much less is understood about how these traits have evolved. Evolutionary shifts in diel niche have occurred numerous times among vertebrates, including in the lineage leading to humans. Moreover, shifts to diurnality from nocturnal ancestors represents one of the critical evolutionary transitions that facilitated mammalian dominance of terrestrial niches after the end of the Mesozoic⁶. Therefore, defining patterns of naturally occurring molecular variation in diel niche traits has the potential to provide fundamental insights into the genomic basis of mammalian diversification.

The African striped mouse, Rhabdomys pumilio (Family Muridae, Subfamily Murinae), a rodent distinguished by its prominent dorsal stripe pattern²⁰, is notable for having transitioned from the ancestrally-nocturnal diel niche of Muridae to a diurnal one (Figure 1A)⁸. African striped mice (hereafter, striped mice/mouse) show several adaptive features associated with a diurnal lifestyle, including a cone-rich retina and a UV-blocking lens^9,10 and have also proven to be a useful model for studying the electrophysiological aspects of circadian timekeeping¹¹. These characteristics, together with the striped mouse’s close phylogenetic relationship with the nocturnal laboratory mouse (Mus musculus)^7,21 present an excellent comparative model system in which to unravel the molecular mechanisms and selective pressures underlying diel niche evolution in mammals. Here, we present a high-quality genome assembly for the striped mouse, which we use to facilitate experimental interrogation of diurnality-associated phenotypes.

Figure 1. The striped mouse reference genome.

(A) Adult striped mouse photographed during the day, showing its characteristic dorsal stripe pattern. (B) Hi-C contact map of the 24 chromosome-scale scaffolds. Enrichment of long-range contacts is shown in red. Higher pixel intensity (red) represents a greater number of contacts between loci. Strong enrichment along the diagonal demonstrates accurate scaffold assembly. (C) Comparison of BUSCO gene recovery in the striped mouse reference genome compared to that of 23 other species in the family Muridae. The striped mouse shows high recovery of single-copy mammalian benchmark orthologs with low rates of duplication and fragmentation. Photography credit: Trevor Hardaker.

See also Figure S1 and Data S1A, B.

Striped mouse genome

To permit functional and comparative genomic studies of diel niche traits in the striped mouse, we first produced a high-quality reference genome assembly for this species, using a combination of linked-read technology and Hi-C scaffolding. The resulting assembly is ~2.3Gb in length with a G+C content of ~41%, comparable to that of other species in its subfamily (Murinae; Data S1A). N50 scaffold length is 79.8Mb, which exceeds that of most available murine genomes and is approximately equal to that of its close relative the African grass rat, Arvicanthis niloticus (Figure S1). Approximately 95% of the assembly is contained within 24 scaffolds having a length greater than or equal to 25Mb with the remainder distributed across smaller scaffolds (Figure 1B and Data S1A). Given the 2n=48 karyotype found in striped mice²², this indicates that the assembly very closely approaches full chromosome-scale, though some chromosomal segments in the lower megabase range remain unplaced. These metrics, together with a low gap percentage (~1%), demonstrate that our assembly exceeds most published murid genome assemblies in terms of contiguity. Next, to assess the assembly’s completeness and integrity, we annotated benchmarking mammalian orthologs using Benchmarking Universal Single-Copy Orthologue (BUSCO)²³. Recovery of complete single-copy BUSCO genes was high at 93.3%, with a few more genes (1.1%) recovered as partial copies and the rate of duplication was low (2.3%; Figure 1C, Data S1B). These values compared favourably with most sequenced murid species (Figure 1C). Next, we sequenced RNA from multiple striped mouse tissues and used these data to produce transcriptome-based gene annotations (see STAR Methods). BUSCO analysis of the 31,928 predicted protein sequences (encoded by 28,594 protein-coding genes) showed that 93.3% of benchmarking genes were present as complete single copies. Taken together, these metrics demonstrate that our striped mouse assembly represents a high-quality resource for genomic studies of this species.

Transcriptomics of circadian rhythm

Circadian clocks are essential for generating and coordinating daily animal rhythms²⁴. These rhythms result from the activity of an intracellular molecular oscillator that operates with a ~24-hour periodicity^25,26. The core molecular clock consists of a transcriptional-translational feedback loop (TTFL) in which Bmal1 and Clock genes constitute a ‘positive arm’, activating the transcription of clock-controlled genes (CCGs). These CCGs include the canonical clock genes Period (Per1-3), Cryptochrome (Cry1-2) and Rev-Erbα and β (which then act as negative regulators of Clock and Bmal1 activity), as well as other ‘output’ genes endowing rhythmicity on physiological processes^26,27. In mammals, the hypothalamic suprachiasmatic nuclei (SCN) contain a master circadian oscillator, which is synchronised to the light:dark cycle and in turn sets the phase of local oscillators in cells and tissues throughout the body²⁷. Diel niche may alter the SCN oscillator’s experience of the synchronizing light:dark cycle as nocturnal animals typically adopt light avoiding strategies during the day. This could in turn impose a more intermittent exposure than is the case for species active through the day. In addition, as the SCN clock adopts a similar phase relationship with the light:dark cycle irrespective of preferred temporal niche in mammals^28–30, it follows that diel niche also alters the phase relationship between the central clock and many overt rhythms.

Given the striped mouse’s evolutionary transition to diurnality, we sought to characterize daily variations of gene expression in the SCN and peripheral tissues to define the features of its circadian organisation. To do this, we first performed RNA-Seq on the SCN, retina, lung and liver of striped mice stably entrained to a 12:12 light-dark cycle, at two time points: 2 hours after lights on (Zeitgeber Time (ZT2); coinciding with high behavioural activity), and 2 hours after lights off (ZT14) during the animal’s resting/sleep stage (Figure 2A). The number of unique transcripts represented in our RNA-Seq datasets ranged from 19,258 for the retina to 14,381 for the SCN (Figure 2B). Of these, substantial fractions showed differential expression between ZT2 and ZT14 in all tissues, making up 23–33% of all expressed genes in retina, lung and liver, and ~7% of the transcripts in the SCN (FDR<0.05) (Figure 2B and C). Functional annotation of differentially expressed genes (DEGs) using the Kyoto Encyclopedia of Genes and Genomes (KEGG) resource revealed that day:night differences extended to numerous pathways (Figure S2), including those involved in core cell activities (including circadian rhythm) as well as more tissue specific functions (e.g., fatty acid biosynthesis).

Figure 2. Rhythmicity of the diurnal African striped mouse transcriptome across central and peripheral tissues.

(A) Daily general activity pattern of striped mice under 12h light: 12h dark cycle. Values are expressed as mean ± SEM (n=10). Grey area indicates period of darkness. Zeitgeber time ZT0 corresponds to time of lights on. Dotted lines indicate timing of tissue collection during the day (ZT2) and at night (ZT14). (B) Number of day:night differentially expressed genes (DEGs; orange) and non-DGE (grey) across different tissues (retina (RET), lung (LUN), liver (LIV), and suprachiasmatic nucleus (SCN)). Numbers on the right indicate the percentage (%) of DEGs within each tissue. (C) Number of down- and up-regulated DEGsat night compared to daytime across tissues. Numbers of genes upregulated at night (log2 fold change (LFC>0) are indicated in orange and those downregulated (LFC<0) in blue. Only genes with an adjusted p-value below 0.05 were considered as DEGs. (D) Heatmaps showing expression levels (normalized values, z-score) for core clock genes (Clock, Bmal1, Npas2, Pers, Crys, Rev-erb, Chrono, Dpb, Tef, Hlf and Nfil3) and representative genes key for specific local-tissue function during the day (ZT2) and at night (ZT14). Squares across rows correspond to individual samples (n=10 day and n=10 night). Gene labels in black indicate DEGs. (E) Radial phase plot showing the distribution of times of peak expression relative to prior light:dark cycle (Time 0 = time of lights on) for genes that exhibited circadian rhythmicity in liver (left) and lung (right) under constant dark conditions (841 and 202 genes, respectively). (F) Circadian profile of Rev-erbα expression across the four tissues (Time 0 = clock time of lights on for prior light:dark cycle). (G) Heatmaps showing expression levels (normalized values, z-score) for core clock genes (Clock, Bmal1, Npas2, Pers, Crys, Rev-erb, Chrono, Dpb, Tef, Hlf and Nfil3) in SCN, RET, LIV and LUN collected over 24h in constant dark conditions (at 1h interval). Gene labels in bold indicate cycling genes with a BH.Q <0.1. Grey columns in the RET (Time 0h) and LUN (Time 17h) heatmaps represent missing data.

See also Figure S2 and Data S1.

Turning to the question of the timing of gene expression with respect to the light:dark cycle, we first interrogated the dataset for a small number of core clock genes (elements of the circadian TTFL and clock output transcription factors) (Figure 2D). In all tissues we found that components of the positive arm, including Bmal1 and Npas2 (paralog of Clock) were higher at night, while most negative regulators (Per, Cry2 and Rev-Erb) were highly expressed during the day. The antiphase arrangement of these two groups of genes is consistent with their role as positive vs negative elements of the TTFL. Similarly, in all tissues, expression of clock-controlled transcription factors Chrono (Circadian-associated transcriptional repressor) and the circadian PAR-domain basic leucine zipper transcription factors Dbp, Tef, and Hlf were higher during the day, while Nifl3 expression was higher at night.

The appearance of similarly phased day/night differences in core clock gene expression across all tissues tested represents a stark difference in daily rhythmicity of striped mice compared to that reported in laboratory mice and rats (Rattus sp.), in which genes from the negative limb of the TTFL are low at night in the SCN and retina, but high at night in peripheral tissues including the liver and lung^31–36. In accordance with that observation, a selection of DEGs considered critical to tissue-specific functions in striped mice had a similar phase to that reported in laboratory mice in the SCN and retina, but not in liver and lung (Figure 2D). Thus, in the SCN, genes associated with neuronal signaling and excitability (Rgs16, Rasd1, and Prok2) adopted a similar phase in striped mice to that previously shown in laboratory mice^37,38. The same was true for photopigment Rho and dopamine signaling (Drd4) genes in the retina^39,40. Interestingly, the expression profile of melanopsin (Opn4) in striped mouse was more similar to rats than to mice^41,42. In the lung, conversely, we found that members of the chemokine family (associated with inflammatory response) were higher at ZT14 in striped mice, vs published work showing a peak in the early light phase in mice^35,43. In the liver, clock gene expression is tightly linked with energy metabolic regulators such as Peroxisome proliferator-activated receptors (PPARs)⁴⁴. In striped mouse liver, Pparα was differentially expressed between day and night, as were genes encoding proteins defining glycogenolysis (Pygl) and glycogenesis (Gys2). In common with our findings in the lung, the phasing of these processes was opposite to that reported for laboratory mice⁴⁵.

Taken together, these transcriptome comparisons between day and night suggest a fundamental change in circadian organisation in striped mice compared to laboratory mice. Elements of the molecular clock and key physiological processes in lung and liver in striped mice have adjusted their phase with respect to the light:dark cycle, to match the diurnal pattern of rhythms in overt behaviour and gross physiology (activity and body temperature). This leaves daily patterns of gene expression in phase across SCN and periphery in striped mice (vs out of phase in laboratory mice).

To extend these observations and determine whether these patterns represented a genuine change in circadian organisation, we turned to a description of rhythms in constant darkness. Based upon recommendations to sample at high density when assessing rhythmic gene expression^46,47 we culled individual striped mice at 1 h intervals over 24 hrs in constant darkness and collected SCN, retina, liver, and lung tissue for RNAseq analysis. We identified an average of 20,297 expressed genes in these samples (17,443 in SCN; 19361 in liver; 21,824 in lung; 22,558 in retina) and applied the ‘R’ package MetaCycle to identify genes with a significant circadian variation in expression. The number of genes defined as rhythmic using this approach was reduced compared to those showing differential expression between ZT2 and 14, likely due to the smaller number of samples per timepoint (Figure 2B and C). Thus, 841 genes were rhythmic at BH.Q < 0.1 in the liver, but this number was smaller for lung, SCN, and retina (202, 20, and 14 genes, respectively). In liver and lung, acrophyses for rhythmic genes covered the circadian cycle, but in common with reports in some other species^29,48, they were especially common at two phases at nearly opposite sides of the cycle (in this case around projected dawn and dusk; Figure 2E). Turning to the phasing of rhythms, only one of the canonical clock genes (Rev-erbα) was significantly rhythmic in all datasets (Figure 2F and G). Consistent with the concept that circadian rhythms have a similar phase across SCN and periphery in striped mice, Rev-erbα expression peaked around the end of the subjective night (subjective time 20–0) in all 4 tissues (Figure 2F). More clock genes showed significant rhythmicity in the lung and liver, and in these cases, elements of the positive limb were higher during the subjective night (Clock, Bmal1, Npas2), while elements of the negative limb were more highly expressed during the subjective day (albeit with slightly phase difference for Cry1). It was harder to detect rhythmicity in the SCN, but all 3 clock genes in this tissue were elements of the negative limb (Rev-erbα, Per1 and Per2) and all were higher during the subjective day consistent with their phasing in lung and liver (Figure 2G). In summary then, our transcriptome analyses suggest that circadian gene expression adopts a similar phase across tissues in striped mice in both light-dark cycle and constant dark. This represents a comprehensive realignment of circadian organisation across the body compared to that described in nocturnal mice and rats^25,33.

Shifts in selective regime associated with the evolution of diurnality

Diurnal temporal niche presents numerous ecological challenges distinct from those experienced by nocturnal species. Therefore, we next sought to use our striped mouse genome assembly to identify genomic signatures of selection that might indicate changing selective regimes in response to these factors. To accomplish this, we first identified orthologous protein-coding transcripts across the genomes of the striped mouse and 24 murid species with publicly available assemblies (Figure 3A and Data S1C). After filtering, we constructed 62,879 alignments of orthologous transcripts, each with representative sequences from a minimum of 10 species, which were used for subsequent analyses. Using these ortholog alignments, we performed analysis of relative evolutionary rates (RERs) to identify genes showing accelerated evolution relative to the per-species background rate (Figure S3, see Methods)⁴⁹.

Figure 3. Selection on phototransduction genes.

(A) A phylogeny of species used in comparative genomic analyses. Blue colour indicates the subfamily Murinae to which the striped mouse (red font) belongs. Green colour indicates the genus Acomys (subfamily Deomyinae) which served as an outgroup. (B) Enrichment of Gene Ontology Biological Process and KEGG pathway terms among striped mouse genes showing the greatest evolutionary acceleration relative to the background rate. Striped mouse accelerated genes are highly enriched for functions related to phototransduction, particularly those related to the rhodopsin cascade. Colour indicates FDR q-value. (C) Plots showing the relative evolutionary rates (RERs) of three core genes involved in rhodopsin-mediated phototransduction: Rhodopsin Kinase (Grk1), Cyclic Nucleotide Gated Channel Subunit Alpha 1 (Cnga1), and Solute Carrier Family 24 Member 1 (Slc24a1). Compared to closely related murids (blue), phototransduction locus orthologs in the striped mouse (red) show a markedly elevated evolutionary rate.

See also Figures S3, S4 and Data S1C–J.

Given the changes in circadian organisation in the striped mouse identified by our transcriptomic survey, we first examined the RERs of genes encoding the core clock components. These revealed no evidence of strong evolutionary acceleration compared to background rates of coding gene evolution (Data S1D). This suggests that shifts in diel niche do not have strong implications for the coding sequences of proteins responsible for generating circadian rhythms. It remains possible that evolution of the clock proceeds mainly via alterations in cis-regulatory elements but overall, this finding is consistent with the view that functional constraints on the molecular machinery of circadian rhythm generation are not altered by switches in diel niche.

Functional annotation revealed that the Gene Ontology and KEGG terms most significantly enriched among accelerated genes in the African striped mouse were related to the rhodopsin-mediated phototransduction cascade (Figure 3B, Data S1E). These terms were comprised of 8 striped mouse accelerated genes (Grk1, Slc24a1, Rho, Sag, Pde6g, Cnga1, Gna11, and Metap2) (Figure 3C and Figure S4). The majority of striped mouse accelerated genes (i.e., Grk1, Slc24a1, Rho, Sag, Pde6g, Cnga1) are expressed specifically in rods, a type of light-sensitive photoreceptor cell found in the retina that specializes in dim-light vision (i.e., scotopic vision)⁵⁰. The functional specialization of rods results from the action of genes that participate in the rod phototransduction cascade, a process by which light is converted into an electrical signal. Among the 8 striped mouse accelerated genes, Rho, Cnga1, and Pde6g play key roles in phototransduction activation whereas Grk1, Slc24a1, and Sag are critical for photoresponse recovery. Interestingly, many of these genes have undergone positive selection in other lineages experiencing evolutionary shifts in light conditions. For example, in a comparison among bird species with different diel niches, Sag, Slc24a1, and Cnga1 showed strong positive selection in owls, a nocturnal bird with a visual system tuned for hunting at night⁵¹. In falcons, which are characterized by their high-speed pursuit of prey, sometimes under relatively low-light conditions, Grk1 and Slc24a1 also showed evidence of positive selection^52,53. As catching prey in flight involves tracking fast-moving objects, having enhanced temporal resolution of vision likely leads to effective hunting. In some cases, the physiological relevance of protein coding changes in dim-light vision genes has been directly established through functional tests⁵⁴. However, while predatory birds may benefit from enhanced low-light vision mediated by rhodopsin phototransduction, it is less clear whether a species transitioning away from low-light conditions like the striped mouse would require such adaptations. Elevated substitution rates in striped mouse rod phototransduction genes may instead be explained by a reduced reliance on rod vision and a corresponding relaxation of purifying selection. Therefore, to address this uncertainty we re-examined striped mouse rod phototransduction genes using the branch-site model, a Ka/Ks-based test for positive selection. Our analysis indicated that acceleration of striped mouse phototransduction genes was not consistent with positive selection (p > 0.01; Data S1F). This suggests that rather than necessitating adaptive changes in rod function, the importance of rod-based vision was reduced following the transition to diurnality in the striped mouse, leading to relaxed purifying selection on rod phototransduction.

The striped mouse belongs to the tribe Arvicanthini which includes several closely related African murines exhibiting degrees of diurnality^7,55,56. This suggests that the initial diel niche shift in the striped mouse lineage occurred in its most recent common ancestor (MRCA) with the other diurnal arvicanthines. This raises the question of whether the observed pattern of relaxed constraint on striped mouse rod phototransduction genes is unique to this species or reflects a common trajectory within its diurnal lineage. As our dataset included the genome assembly for the diurnal African grass rat (A. niloticus), an arvicanthine species closely related to striped mice (Figure 3A), we next examined patterns of sequence evolution in this species as well as its common ancestor with the striped mouse. In stark contrast to the striped mouse, rod phototransduction genes were not found to be enriched among African grass rat accelerated genes, with only Sag showing a strongly elevated relative evolutionary rate (Data S1G–H). Branch-site tests indicated that, as in the striped mouse, acceleration of Sag in the grass rat was consistent with relaxed purifying selection (Data S1F). In the inferred MRCA of the striped mouse and grass rat, rod phototransduction genes showed a more similar pattern to that observed in the striped mouse, with a similar complement of genes in this pathway exhibiting elevated evolutionary rates (Data S1I–J). Overlapping genes included Grk1, Slc24a1, Pd36g, Rho, Gna11 and Sag. A further accelerated gene unique to the ancestor, Gnat2, was also identified (Data S1I). Interestingly, while branch-site tests generally rejected adaptation in most of these ancestral genes, evidence of positive selection was identified in Sag, suggesting that this gene may have experienced and early phase of adaptation during the transition to diurnality (Data S1F).

Taken together, our results indicate that the transition to diurnality in the African striped mouse primarily corresponds with broad relaxation of purifying selection on rod phototransduction genes, likely reflecting an overall reduced requirement for acute low-light vision. Moreover, our examination of the related African grass rat as well as their common ancestor provides molecular evidence supporting an ancestral origin for diurnality prior to the striped mouse-grass rat divergence. The initial shift toward a diurnal niche in this lineage was accompanied by widespread relaxation of selection on rod phototransduction genes, as well as adaptation in a small number of sites in the Sag gene, encoding S-arrestin. Interestingly, our findings also indicate that the selective regimes experienced by African grass rats and striped mice since their divergence from a common ancestor have not been uniform, with the African grass rat lacking the extended period of relaxed selection seen in the striped mouse. While the causes of this difference remain unclear, one possibility is that the striped mouse genus Rhabdomys has proceeded further in its diel niche transition than the grass rat genus Arvicanthis. The presence of dorsal pigment patterns with potential camouflaging effects in Rhabdomys (absent in Arvicanthis) may support this notion⁵⁷, as do reports that Arvicanthis can readily adopt nocturnality in the laboratory^56,58. Interestingly, this pattern of evolutionary acceleration differs markedly with that of circadian clock genes. In contrast to clock genes which function in many tissues across the body, expression of the rhodopsin phototransduction cascade is restricted to rod cells of the retina. Differing patterns of pleiotropy potentially create a scenario under which changes in circadian organization may favour cis-regulatory evolution while coding sequence evolution may be permitted for genes in the rhodopsin cascade.

Dim-light vision in striped mice

Relaxed purifying selection on rod phototransduction genes in the striped mouse suggests that there may be corresponding functional changes in dim-light vision compared to nocturnal relatives. To test this prediction, we applied electroretinography to compare visual performance under scotopic in striped mice vs nocturnal laboratory mice. Animals were dark adapted for 6 hours and exposed to full field, square wave modulations in light intensity (80.5% Michelson contrast) against a dim background light (9.6 log photons/cm²/s, equivalent to 3.58 R*/rod/s) and across a range of frequencies (1–25Hz). Both laboratory and striped mice responded to elements of this stimulus as evidenced by modulations in the electroretinogram (ERG) at the appropriate frequency (Figure 4A). To facilitate comparisons between species, we applied fast Fourier transform (FFT) to objectively quantify ERG oscillation amplitudes occurring at the stimulus frequency (Figure 4B). In both species, FFT amplitude fell away as temporal frequency increased. However, there was a species difference in this relationship (F- test for sigmoidal curve fit p=0.006) with the reduction in amplitude occurring at lower frequencies in striped mice (EC50=2.0 vs 9.5 Hz in striped vs laboratory mice). This deficit in striped mouse response to higher frequency flicker was specific to dim light conditions, as when we presented this same stimulus under a brighter background light we found that in fact striped mouse ERGs had higher FFT amplitude than mice across the higher frequencies (Figure 4C and D). Further work will be required to determine whether there is a mechanistic link between the accelerated evolution of rod phototransduction genes and this change in dim light visual response. The genes showing accelerated evolution encode proteins which collectively define the gain and longevity of the rod photoresponse, raising the possibility of a causative link between accumulated changes in these proteins during striped mouse evolution and reductions in temporal resolution for rod vision. Testing that possibility will require direct recordings of the striped mouse rod response and, ideally, expressing striped mouse proteins in the rod of laboratory mouse. Whether or not such a direct causative relationship exists, the discovery of reduced performance in dim-light vision in striped mice is consistent with a reduction in the selective constraints on rod photoreception in this species.

Figure 4. ERG responses in African striped mice and Laboratory mice.

(A) Mean ERG traces from laboratory mice (left) and striped mice (right) in response to square-wave modulations (80.5% Michelson contrast) presented against a dim (9.6 log photons/cm²/s) background. Frequency in Hz shown to left, and light modulation shown below each trace. Scale bar = 100ms (x-axis), 50μV (y axis). (B) Mean±SEM FFT amplitude at modulation frequency for square-wave modulation across frequency range at 9.6 log photons/cm²/s background for laboratory mice (black) and striped mice (red). Data for each species are fitted with a separate sigmoidal curve (F test comparison p<0.001) with variable slope. C and D as A and B but for stimuli presented against a bright background (14.6 log photons/cm²/s). Scale bar = 100ms (x-axis), 25μV or 250μV (y axis) for laboratory and striped mice respectively. n=5 per group.

By taking advantage of naturally evolved, adaptive phenotypes in African striped mice, our work provides insights into the transcriptional changes and genomic patterns of natural selection associated with temporal niche transition in a diurnal rodent. We identify a district pattern of molecular clock reorganization concomitant with the inverted diel activity in this species and raise important new questions about show such phasing is achieved at the regulatory level. Moreover, we find evidence of relaxed selection on rod phototransduction, reflecting a reduced reliance on low-light vision, supported by in vivo retinal measurements. More broadly, our approach exemplifies how integrating interdisciplinary approaches, including whole-genome sequencing, transcriptomics, comparative genomics, and functional experiments, can improve our understanding of the evolutionary processes shaping organismal traits in the wild.

STAR Methods

Resource availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Ricardo Mallarino (rmallarino@princeton.edu).

Materials availability

This study did not generate new unique reagents

Data and code availability

RNA-Seq data have been deposited in the ArrayExpress database at EMBL-EBI under accession number E-MTAB-12024. The genome assembly, 10X Chromium Linked Reads, and Dovetail Hi-C reads have been deposited in the BioProject database: PRJNA858857. Lastly, we have created a FigShare repository (https://doi.org/10.26188/20321655) containing the following: 1) Rhabdomys_pumilio_final.gff.gz: A raw GFF-formatted gene annotation set for the Rhabdomys pumilio genome produced by Funannotate and used in transcriptomic analyses; 2) Rhabdomys_pumilio.mouse_gene_name_final.gff.gz: A copy of the above GFF-formatted gene annotation set for the Rhabdomys pumilio genome, in which gene symbols from the laboratory mouse (Mus musculus) have been assigned to their predicted R. pumilio orthologs; 3) CompGenAnno.tar.gz: A folder of GFF-formatted annotations used in comparative genomic analyses, produced by directly lifting-over gene annotations from the M. musculus genome (annotation: GCF_000001635.27_GRCm39_genomic.gff, assembly: GCF_000001635.27_GRCm39_genomic.fna) onto each of 23 other murid genome assemblies. Additionally, a manifest of each reference genome can be found in .tsv format (manifest_of_genome_assemblies_and_liftover_annotations.txt) along with a file with locus trees (locus_trees.txt) for each orthologous group of genes used in comparative genomic analyses are provided; 4) Table_of_RER_data_for_examined_species.xlsx: A large table showing relative evolutionary rates measurements for each orthologous group of gene sequences (referenced against a M. musculus transcript), for each species examined. Each species may be listed in multiple columns, reflecting different species representation for each orthologous group (i.e., representing cases in which the branch to a given leaf node originates at a different ancestral node due to sister species not represented in that alignment); and 5) Rhabdomys_pumilio_Princeton_asm1.0_preNCBI.fasta.gz: A copy of the genome assembly prior to any re-formatting that NCBI performs after upload.

This paper does not report original code.

Any additional information required to reanalyse the data reported in this paper is available from the lead contact upon request.

Experimental model

All experiments done at the University of Manchester were performed in accordance with the UK Animals, Scientific Procedures Act of 1986, and the study was approved by the UK Home Office under the animal license PP3176367. All experiments done at Princeton University were performed with authorisation from Princeton University’s IACUC.

Rhabdomys pumilio:

F10 descendants of wild-derived striped mice (R. pumilio, originating from Goegap Nature Reserve, South Africa, S 29° 41.56′, E 18° 1.60′) were obtained from a captive colony at the University of Zurich (Switzerland) and are maintained at Manchester University and Princeton University. Striped mice are kept at a 12:12 light-dark cycle and given food ad libitum. Both males and female adult striped mice were used for this study.

Mus musculus:

All laboratory mice used in this study were from the C57BL/6 inbred strain, obtained from (Envigo). Importantly, while many inbred laboratory mice harbor mutations in their visual systems that may alter functional measurements, the strain used his has a fully functional visual system. Both males and female adult laboratory mice were used for this study

Methods Details

Genome assembly and annotation

We collected 50mg of thigh muscle tissue from a euthanized female African striped mouse from the captive colony housed at Princeton University for genome sequencing. High molecular weight DNA was extracted with the Qiagen MagAttract HMW kit (NEB 67563), quantified with the Qubit DNA HS kit (Thermofisher Q32851) and DNA fragment size distribution was visualized using the BioAnalyzer 2100 High Sensitivity chip. We next prepared a single genome library using the 10X Genomics Chromium system with the Genome v2 Library Prep kit and performed sequencing on an Illumina HiSeq 2500 Rapid Flowcell in 2×150bp format. In total, ~275Gb of sequence data were generated, corresponding to approximately 119X haploid genome coverage. A primary assembly was generated using the Supernova2 pipeline (10X Genomics). Next, additional tissue samples were provided to Dovetail Genomics who generated Hi-C libraries, which were used to scaffold the primary assembly with the HiRise pipeline. Genome assembly metrics were calculated using the stats.sh script included in the BBMap package v37.93⁵⁹. BUSCO v5.2.2²³ was used to annotate benchmarking mammalian orthologs from the mammalia_odb10 database across murid genome assemblies.

Gene prediction and annotation of the assembled genome was carried out using Funannotate v1.8.1⁶⁰. RNA data generated from peripheral tissues (see below) was used as evidence for gene prediction. Prior to gene prediction, RNA data was quality and adapter trimmed using TrimGalore v0.6.6⁶¹. Repeat models were generated using Red⁶². These repeats were used as BLASTX queries against the UniProt/Swiss-Prot database of curated proteins. Sequences that significantly matched with any protein and the flanking 50bp were removed from the repeat models using ProtExcluder⁶³. Filtered repeat models were applied to the genome using RepeatMasker⁶⁴. BUSCO was used to estimate annotation completeness (93.3% complete single copy, 1.1% recovered as partial copies).

General activity monitoring

General activity rhythms were recorded for 2 weeks in single housed striped mice kept under a 12:12 light:dark cycle (Zeitgeber Time (ZT) 0 corresponds to the time of lights on, and ZT12 to lights off) using a passive infrared motion sensor system, as previously described. Data were acquired in 10s bins and expressed as % of the ceiling activity reading of these sensors in that time window, in this way we calculate activity as % of the maximum amount of activity detectable with the sensor⁶⁵. Daily 24h-profile of general activity was calculated for each animal using 30min binned data across a period of 7 days and averaged from a group of 10 animals.

RNA extraction, library preparation and sequencing

ZT2 and ZT14 analysis

Tissue samples from animals (age ~13mo; male and female) kept under 12:12 light:dark cycle were collected at ZT2 and ZT14 (10 animals per time point). Animals were sedated with isoflurane and immediately culled by cervical dislocation.

Retina, lung, and liver samples:

Total RNA was extracted from < 5 mg frozen tissue using the Qiagen RNeasy Mini kit (Qiagen 74104) according to the manufacturer’s instructions. Total RNA was submitted to the Genomic Technologies Core Facility (GTCF, University of Manchester, UK). Quality and integrity of the RNA samples were assessed using a 4200 TapeStation (Agilent Technologies) and then libraries generated using the Illumina^® Stranded mRNA Prep. Ligation kit (Illumina, Inc. 20040532) according to the manufacturer’s protocol. Briefly, total RNA (typically 1ug) was used as input material from which polyadenylated mRNA was purified using poly-T, oligo-attached, magnetic beads. Next, the mRNA was fragmented under elevated temperature and then reverse transcribed into first strand cDNA using random hexamer primers and in the presence of Actinomycin D (thus improving strand specificity whilst mitigating spurious DNA-dependent synthesis). Following removal of the template RNA, second strand cDNA was then synthesized to yield blunt-ended, double-stranded cDNA fragments. Strand specificity was maintained by the incorporation of deoxyuridine triphosphate (dUTP) in place of dTTP to quench the second strand during subsequent amplification. Following a single adenine (A) base addition, adapters with a corresponding, complementary thymine (T) overhang were ligated to the cDNA fragments. Pre-index anchors were then ligated to the ends of the double-stranded cDNA fragments to prepare them for dual indexing. A subsequent PCR amplification step was then used to add the index adapter sequences to create the final cDNA library. The adapter indices enabled the multiplexing of the libraries, which were pooled prior to cluster generation using a cBot instrument. The loaded flow-cell was then paired-end sequenced (76 + 76 cycles, plus indices) on an Illumina HiSeq4000 instrument. Finally, the output data was demultiplexed and BCL-to-Fastq conversion performed using Illumina’s bcl2fastq software, version 2.20.0.422.

Suprachiasmatic nucleus (SCN) samples:

Punches were collected from 1 mm thick brain slices containing the SCN using a sample corer. RNA was extracted from snap-frozen brain punches using the Qiagen RNeasy Micro kit (Qiagen 74004) according to the manufacturer’s instructions and quantified using a 4200 TapeStation (Agilent Technologies). Three nanograms of RNA (RIN ≥ 8.7) was used as input material and libraries were prepared by following the SMARTer Stranded Total RNA-Seq Kit V3 - Pico Input Mammalian (Takara Bio 634485) user manual. In brief, samples were fragmented at 94 °C for 4 min prior to first-strand synthesis. Illumina adaptors and indexes were added to single-stranded cDNA via 5 cycles of PCR. Libraries were hybridized to R-probes for fragments originating from ribosomal RNA to be cleaved by ZapR. The resulting ribo-depleted library fragments were amplified with 15 cycles of PCR. Samples were pooled and sequenced as described above.

24-hr circadian transcriptomic profiling

For circadian transcriptome profiling, animals (age ~ 3mo; males and females) were housed in constant darkness, with tissue samples were taken at hourly intervals (one animal per time point) starting at circadian time 18 (6 hours after lights went off for the last time) and continuing for a further 23-hour. Samples were snap frozen and stored at −80C until RNA extraction.

Retina, lung, and liver samples:

Total RNA was extracted from 5–10mg of snap frozen tissue using the Qiagen RNeasy Fibrous Tissue Mini Kit (Qiagen 74704) according to the manufacturer’s instructions. RNA concentration was measured via Qubit 2.0 (ThermoFisher) and RNA quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). mRNA was isolated using the NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs E7490) and libraries were prepped using the NEBNext Ultra II Directional RNA library prep kit for Illumina (New England Biolabs E7760) and NEBNext Multiplex Oligos for Illumina (New England Biolabs E6440), following the manufacturer’s instructions. In brief, mRNA was isolated using OligodT beads and fragmented at 94°C for 15 minutes. cDNA was synthesized, adaptors were added, and the adaptor-ligated DNA was amplified with unique barcodes using 10–12 PCR cycles. Samples were combined into a single pool and sequenced on an Illumina NovaSeq S1 100nt Flowcell (2 × 69-bp format). To increase read depth, a second round of sequencing was performed on an Illumina NovaSeq SP 100nt Flowcell (2 × 69-bp format). Raw reads from both sequencing runs were concatenated prior to downstream processing/analyses.

Suprachiasmatic nucleus (SCN) samples:

Punches were collected from 1 mm thick brain slices containing the SCN using a sample corer. RNA was extracted from snap-frozen brain punches using the Qiagen RNeasy Micro kit (Qiagen 74004) according to the manufacturer’s instructions. RNA concentration was measured via Qubit 2.0 (ThermoFisher) and RNA quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). Libraries were prepared using the SMARTer Stranded Total RNA-Seq Kit V3 - Pico Input Mammalian (Takara Bio 634485) following the manufacturer’s instructions, including the RNA fragmentation step. In brief, samples were fragmented at 94 °C for 4 min prior to first-strand synthesis. Illumina adaptors and barcodes were added to single-stranded cDNA (5 PCR cycles), and ribosomal cDNA was depleted using ZapR v3 and R-Probes v3. Libraries were then amplified using 12 cycles of PCR. Samples were pooled with the retina, lung, and liver samples and sequenced as described above for retina, liver, and lung samples.

Quantification and Statistical Analysis

Differential gene expression analysis:

The quality of the stranded paired-end RNA-Seq reads was assessed using FastQC (v0.11.3)⁶⁶ and FastQ Screen (v0.14.0)⁶⁷, and adapter and low-quality base were trimmed using BBDuk (BBMap suite v38.93)⁵⁹. Processed reads were then mapped against the Rhabdomys pumilio genome and in-house gene annotation (available from the cited FigShare repository) using STAR (v2.7.9a)⁶⁸. The “--quantMode GeneCounts” option was used to obtain read counts per gene from STAR.

In the R environment (v3.6.3), differential gene expression analysis was performed using Bioconductor package DESeq2 (v1.26.0)⁶⁹, and Benjamini-Hochberg (BH) procedure was used to controls false discovery rate (FDR) at 5%. Additionally, the lfcShrink function (with apeglm method) was applied to generate a more accurate log2 fold change estimate. The counts function (with normalized = TRUE) was used to generate normalised counts by the median-of-ratios method as described in ref⁷⁰.

Pathway enrichment analysis:

Pathway enrichment analysis of DEGs was performed using the R package enrichR (v3.0)⁷¹. The KEGG 2019 Mouse gene-set library defined by the Enrichr web server⁷¹ was chosen for gene ontology analysis. Pathway results were ranked by the combined scores (calculated as log(P-value) multiplied by z-score) and pathways with an FDR-adjusted p-value < 0.05 were considered significant.

Rhythmic analysis:

The time-course RNA-seq datasets from SCN, retina, lung and liver were processed using BBDuk and STAR as described above with the exception that the transcript alignments were obtained by using “--quantMode TranscriptomeSAM” when running STAR. RSEM (v1.3.1)⁷² was used to perform gene expression quantification. The transcriptomic data from all four tissues were checked for the presence of outliers using principal component analysis. Two samples, ZT0 from retina and ZT17 from lung, were excluded from the circadian rhythmicity analysis on this basis. Furthermore, genes that were not expressed in any samples were also excluded.

The gene-level transcripts per million mapped reads (TPM) values were used as input to detect rhythmic signals using MetaCycle (v1.2.0)⁷³ in R environment. The “meta2d” function was run with the following parameters: minimum period length of 20, maximum period length of 28, expected period length of 24, phase adjusted with predicted period length, ARSER, JTK_CYCLE and Lomb-Scargle algorithms were used (ARSER was disabled automatically in lung and retina when there were missing values) and the Fisher’s method was used to integrate p-values from the chosen methods. Rhythmic genes were identified with statistical significance BH q-values (meta2d_BH.Q) < 0.1.

Analysis of relative evolutionary rates and positive selection

To permit comparative analysis of relative evolutionary rates, we use the following procedure to predict orthologous genes across 24 murid species. We first downloaded the reference genome for the laboratory mouse M. musculus (GRCm39) and its corresponding RefSeq gene annotations⁷⁴. The mouse genome is of excellent quality and is rigorously annotated using transcriptome data from numerous tissues, thus it was used as the reference species for evolutionary analyses. To extract orthologous transcripts suitable for comparative evolutionary analysis from other murids, we used liftoff v1.6.1⁷⁵ to lift-over gene models from M. musculus to a total of 23 other murid species including the African striped mouse. 17 additional species within the subfamily Murinae were examined, as this group includes the African striped mouse and its diurnal sister genus Arvicanthis, as well as the reference species M. musculus. 5 additional species belonging to the genus Acomys (subfamily Deomyinae) were included as an outgroup (Data S1C).

Next, coding sequences (CDS) were extracted from each murid genome using gffread v0.12.4 and collated into orthologous groups in multifasta files⁷⁶. Each orthologous group was first filtered to remove sequences differing in total length from the M. musculus reference gene by more than 10%. Next, orthologous groups were filtered such that only those with a representative sequence for the African striped mouse and at least one Acomys outgroup species were retained. Next, files in which fewer than 75% of sequences began with an ATG were excluded, as were files containing sequences from fewer than 10 total species. Next, orthologous sequences were aligned using mafft v7.407 with parameters --localpair and --maxiterate 1000⁷⁷. Resulting ortholog alignments were then re-filtered by the criteria above to remove low-quality orthologs and then re-aligned.

To infer a species tree topology for subsequent evolutionary analyses, alignments containing representative orthologs for all 24 murid species were filtered to retain only one representative transcript per gene based on highest alignment score, were concatenated and were then provided to RAxML v8.2.12⁷⁸ with parameters: -f a -x 58744 -p 58744 -# 1000 -m GTRGAMMA, flagging the genus Acomys the outgroup (parameter -o) and using 1000 bootstrap replicates (parameter -# 1000). The topology of the best maximum likelihood tree was found to be consistent with a recent, large-scale phylogeny of Muroids²¹ and was thus taken as the species tree for subsequent evolutionary analyses. RAxML was used to estimate branch lengths on the fixed species tree topology for each ortholog alignment with parameter -f e and -m GTRGAMMA. We chose to use RAxML because it is a commonly used and computationally efficient tool to generate branch length estimates from multiple sequence alignments, which are required as input for our downstream analyses using RER (see below).

Several approaches exist to define evolutionarily accelerated loci. For instance, PhyloP implements multiple tests for evolutionary acceleration relative to a background model of neutral evolutionary rates, usually calculated from 4-fold degenerate sites. The growing recognition that selection is ubiquitous across the genome and that phenomena such as linked-selection can markedly alter the evolutionary rates of 4-fold degenerate sites⁷⁹ calls into question the degree to which such models accurately reflect neutral rates. Moreover, when examining evolution of protein-coding sequences, neutral evolutionary rates may not represent an appropriate null model. Protein-coding genes are generally considered to experience evolutionary constraint as a product of purifying selection. Thus, their expected rate of sequence evolution should be slower than would be expected under neutrality. Given this, we took an alternate approach, by calculated relative evolutionary rates for each sequence in each ortholog alignment using the getAllResiduals function (parameters: transform = “sqrt”, weighted = T, scale = T) implemented in the package RERConverge v0.3.0⁴⁹. Rather than comparing evolutionary rates against a neutral model, RERConverge generates a consensus or “average” tree from a set of input locus trees (i.e., those calculated by RAxML above). Each locus tree is then scaled to the master tree, and residual branch lengths are taken as the relative evolutionary rate (RER) for each ortholog with the locus tree. In this way, shifts in evolutionary rate are relative to a background expectation based on the class of elements being tested. Moreover, given the close evolutionary relationships and similar genome compositions of species being examined in our analyses, plots of ortholog RERs (such as those shown in Figure 3C and Figure S4) can be used to rule out alternate explanations for acceleration based on genomic region (such as GC-biased gene conversion⁸⁰), as acceleration driven by such effects would be present across orthologs from most or all species for a given locus with relative uniformity. After calculating RERs for each group of orthologs, we defined highly accelerated sequences in the African striped mouse as the 95^th percentile of genes with a positive relative evolutionary rate (Figure S3) and focused on these in downstream examinations. It is worth noting that, while the branch lengths for the initial gene trees calculated for each group of orthologous loci represent mean substitutions per site, after scaling the gene tree for comparison with the master/consensus tree, that is no longer strictly the case. Therefore, the RER value is best understood at being unitless.

Highly accelerated genes were examined for enrichment of KEGG Pathway and GO Biological Process terms using the Enrichr web server⁸¹. Dot plots of per-species, per-gene relative evolutionary rates were generated using the foreground2Paths and plotRers functions in RERconverge. Results from the RERconverge analysis for striped mice are listed in Data S1D.

The branch-site model was used to test for evidence of positive selection. Briefly, sequence files for rhodopsin cascade genes were first re-aligned using MACSE v2.06 and branch-lengths recalculated with RAxML (parameter -f e and -m GTRGAMMA)⁷⁸. These alignments and trees were then provided to EasyCodeML v1.0⁸² to perform branch-site tests on the striped mouse, grass rat, and their most recent common ancestor. Multiple testing correction was performed using the Benjamini-Hochberg procedure⁸³.

ERG measurements

ERGs were recorded in adult M. musculus (C57/BL6, aged 4–5 months) and adult male R. pumilio (aged 7–8 months). All ERGs were recorded at subjective midday following dark adaptation for a period of 6 hours. Animals were anaesthetised under isoflurane in a 95/5% Oxygen/CO2 mix at a flow rate of 0.5 – 1.0L/min. Isoflurane concentrations of 5% and 1.5–3.5% were used for induction and maintenance of anaesthesia, respectively. A topical mydriatic (tropicamide 1%; Bausch and Lomb) and hypromellose eye drops were applied to the recording eye before placement of a corneal contact-lens–type electrode. A needle reference electrode (Ambu, Neuroline) was inserted approximately 5mm from the base of the contralateral eye, and a second subcutaneous needle in the scruff acted as a ground. Electrodes were connected to a Windows PC via a signal conditioner (Model 1902 Mark III, CED) that differentially amplified (X3000) and filtered (band-pass filter cut off 0.5 to 200Hz) the signal, and a digitizer (Model 1401, CED). Core body temperature was maintained at 37°C throughout recording with a homeothermic heat mat (Harvard Apparatus).

Visual stimuli were generated with a combination of violet, blue and cyan elements of a multispectral LED light source (Lumencor). Intensities were modulated via pulse width modulations via an Arduino Uno. Light from the light engine passed through a filter-wheel containing neutral-density filters (reducing the light by between 10¹ and 10⁵) and focused onto opal diffusing glass (5mm diameter; Edmund Optics Inc.) positioned <5mm from the eye. All LED intensities were controlled dynamically with a PC. Stimuli were measured at the corneal plane using a spectroradiometer (SpectroCAL II, Cambridge Research Systems, UK) between 350–700nm. Stimuli were presented either as square-wave modulations from background at 80.5% Michelson contrast across a range of frequencies (1, 2, 4, 6, 12, 16, 25 Hz) and at two backgrounds (9.6, 14.6 log photons/cm²/s). ERG responses were analysed in MATLAB (release 2018a, Mathworks), using a fast Fourier Transform (FFT) to analyse ERG responses in the frequency domain. This allowed us to use an objective measure to quantify variations in the ERG signal at the frequency of the visual stimulus presented.

Supplementary Material

1. Data S1. Data supporting main figures of the text, related to Figures 1, 2, and 3.

(A) African striped mouse genome assembly metrics; (B) African striped mouse BUSCO recovery; (C) Genomes accessed in the present study; (D) Evolutionarily accelerated African striped mouse genes; (E) Summary of KEGG and GO Biological Process terms enriched among African striped mouse (Rhabdomys pumilio) accelerated genes; (F) Branch-site tests of positive selection in Arvicanthine accelerated genes; (G) Evolutionarily accelerated genes in the African grass rat (Arvicanthis niloticus); (H) Summary of KEGG and GO Biological Process terms enriched among African grass rat (Arvicanthis niloticus) accelerated genes; (I) Evolutionarily accelerated genes in the R. pumilio and A. niloticus most recent common ancestor; (J) Summary of KEGG and GO Biological Process terms enriched among accelerated genes in the R. pumilio and A. niloticus most recent common ancestor.

Key resources table.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Bacterial and virus strains
Biological samples
Chemicals, peptides, and recombinant proteins
Critical commercial assays
Deposited data
RNA Sequencing Data	This paper	ArrayExpress (EMBL-EBI) under accession number E-MTAB-12024
The genome assembly, 10X Chromium Linked Reads, and Dovetail Hi-C reads	This paper	BioProject database: PRJNA858857
Rhabdomys_pumilio_final.gff.gz: A raw GFF-formatted gene annotation set for the Rhabdomys pumilio genome produced by Funannotate and used in transcriptomic analyses	This paper	FigShare repository (https://doi.org/10.26188/20321655
Rhabdomys_pumilio.mouse_gene_name_final.gff.gz: A copy of the above GFF-formatted gene annotation set for the Rhabdomys pumilio genome, in which gene symbols from the laboratory mouse (Mus musculus) have been assigned to their predicted R. pumilio orthologs.	This paper	FigShare repository (https://doi.org/10.26188/20321655
CompGenAnno.tar.gz: A folder of GFF-formatted annotations used in comparative genomic analyses, produced by directly lifting-over gene annotations from the M. musculus genome (annotation: GCF_000001635.27_GRCm39_genomic.gff, assembly: GCF_000001635.27_GRCm39_genomic.fna) onto each of 23 other murid genome assemblies. Additionally, a manifest of each reference genome can be found in .tsv format (manifest_of_genome_assemblies_and_liftover_annotations.txt) along with a file with locus trees (locus_trees.txt) for each orthologous group of genes used in comparative genomic analyses are provided.	This paper	FigShare repository (https://doi.org/10.26188/20321655
Table_of_RER_data_for_examined_species.xlsx: A large table showing relative evolutionary rates measurements for each orthologous group of gene sequences (referenced against a M. musculus transcript), for each species examined. Each species may be listed in multiple columns, reflecting different species representation for each orthologous group (i.e., representing cases in which the branch to a given leaf node originates at a different ancestral node due to sister species not represented in that alignment).	This paper	FigShare repository (https://doi.org/10.26188/20321655
Rhabdomys_pumilio_Princeton_asm1.0_preNCBI.fasta.gz: A copy of the genome assembly prior to any re-formatting that NCBI performs after upload	This paper	FigShare repository (https://doi.org/10.26188/20321655
Experimental models: Cell lines
Experimental models: Organisms/strains
Rhabdomys pumilio	Mallarino and Lucas laboratory	F10 descendants of wild-derived striped mice (R. pumilio, originating from Goegap Nature Reserve, South Africa, S 29° 41.56′, E 18° 1.60′) were obtained from a captive colony at the University of Zurich (Switzerland) and are maintained at Manchester University and Princeton University
Mus musculus	Lucas laboratory	C57BL/6 inbred strain, obtained from (Envigo)
Oligonucleotides
Recombinant DNA
Software and algorithms
BBMap suite v38.93	Buschnell et al.56	https://sourceforge.net/projects/bbmap/
BUSCO v5.2.2	Seppey et al.22	https://mybiosoftware.com/busco-assessing-genome-assembly-and-annotation-completeness-with-single-copy-orthologs.html
Funannotate v1.8.1	Palmer57	https://github.com/nextgenusfs/funannotate/tree/v1.8.1
TrimGalore v0.6.6	Kruger58	https://github.com/FelixKrueger/TrimGalore
ProtExcluder	Campbell et al.60	https://github.com/NBISweden/ProtExcluder
RepeatMasker	Smit et al.61	https://www.repeatmasker.org/
FastQC (v0.11.3)	Andrews63	https://www.bioinformatics.babraham.ac.uk/projects/fastqc/
FastQ Screen (v0.14.0)	Wingett and Andrews64	https://www.bioinformatics.babraham.ac.uk/projects/fastq_screen/_build/html/index.html
STAR (v2.7.9a)	Dobin et al.65	https://github.com/alexdobin/STAR
DESeq2 (v1.26.0)	Lewis et al.66	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
enrichR (v3.0)	Chen et al.68	https://cran.r-project.org/web/packages/enrichR/index.html
RSEM (v1.3.1)	Li and Dewey70	https://deweylab.github.io/RSEM/
MetaCycle (v1.2.0)	Wu et al.72	https://github.com/gangwug/MetaCycle
liftoff v1.6.1	Shumate and Salzberg71	https://github.com/agshumate/Liftoff
RAxML v8.2.12	Stamakis74	https://cme.h-its.org/exelixis/web/software/raxml/
RERConverge v0.3.0	Pertea and Pertea75	https://github.com/nclark-lab/RERconverge
EasyCodeML v1.0	Gao et al.79	https://github.com/BioEasy/EasyCodeML
Other

LIFE SCIENCE TABLE WITH EXAMPLES FOR AUTHOR REFERENCE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit monoclonal anti-Snail	Cell Signaling Technology	Cat#3879S; RRID: AB_2255011
Mouse monoclonal anti-Tubulin (clone DM1A)	Sigma-Aldrich	Cat#T9026; RRID: AB_477593
Rabbit polyclonal anti-BMAL1	This paper	N/A
Bacterial and virus strains
pAAV-hSyn-DIO-hM3D(Gq)-mCherry	Krashes et al.1	Addgene AAV5; 44361-AAV5
AAV5-EF1a-DIO-hChR2(H134R)-EYFP	Hope Center Viral Vectors Core	N/A
Cowpox virus Brighton Red	BEI Resources	NR-88
Zika-SMGC-1, GENBANK: KX266255	Isolated from patient (Wang et al.2)	N/A
Staphylococcus aureus	ATCC	ATCC 29213
Streptococcus pyogenes: M1 serotype strain: strain SF370; M1 GAS	ATCC	ATCC 700294
Biological samples
Healthy adult BA9 brain tissue	University of Maryland Brain & Tissue Bank; http://medschool.umaryland.edu/btbank/	Cat#UMB1455
Human hippocampal brain blocks	New York Brain Bank	http://nybb.hs.columbia.edu/
Patient-derived xenografts (PDX)	Children’s Oncology Group Cell Culture and Xenograft Repository	http://cogcell.org/
Chemicals, peptides, and recombinant proteins
MK-2206 AKT inhibitor	Selleck Chemicals	S1078; CAS: 1032350–13-2
SB-505124	Sigma-Aldrich	S4696; CAS: 694433–59-5 (free base)
Picrotoxin	Sigma-Aldrich	P1675; CAS: 124–87-8
Human TGF-β	R&D	240-B; GenPept: P01137
Activated S6K1	Millipore	Cat#14–486
GST-BMAL1	Novus	Cat#H00000406-P01
Critical commercial assays
EasyTag EXPRESS 35S Protein Labeling Kit	PerkinElmer	NEG772014MC
CaspaseGlo 3/7	Promega	G8090
TruSeq ChIP Sample Prep Kit	Illumina	IP-202–1012
Deposited data
Raw and analyzed data	This paper	GEO: GSE63473
B-RAF RBD (apo) structure	This paper	PDB: 5J17
Human reference genome NCBI build 37, GRCh37	Genome Reference Consortium	http://www.ncbi.nlm.nih.gov/projects/genome/assembly/grc/human/
Nanog STILT inference	This paper; Mendeley Data	http://dx.doi.org/10.17632/wx6s4mj7s8.2
Affinity-based mass spectrometry performed with 57 genes	This paper; Mendeley Data	Table S8; http://dx.doi.org/10.17632/5hvpvspw82.1
Experimental models: Cell lines
Hamster: CHO cells	ATCC	CRL-11268
D. melanogaster: Cell line S2: S2-DRSC	Laboratory of Norbert Perrimon	FlyBase: FBtc0000181
Human: Passage 40 H9 ES cells	MSKCC stem cell core facility	N/A
Human: HUES 8 hESC line (NIH approval number NIHhESC-09–0021)	HSCI iPS Core	hES Cell Line: HUES-8
Experimental models: Organisms/strains
C. elegans: Strain BC4011: srl-1(s2500) II; dpy-18(e364) III; unc-46(e177)rol-3(s1040) V.	Caenorhabditis Genetics Center	WB Strain: BC4011; WormBase: WBVar00241916
D. melanogaster: RNAi of Sxl: y[1] sc[*] v[1]; P{TRiP.HMS00609}attP2	Bloomington Drosophila Stock Center	BDSC:34393; FlyBase: FBtp0064874
S. cerevisiae: Strain background: W303	ATCC	ATTC: 208353
Mouse: R6/2: B6CBA-Tg(HDexon1)62Gpb/3J	The Jackson Laboratory	JAX: 006494
Mouse: OXTRfl/fl: B6.129(SJL)-Oxtrtm1.1Wsy/J	The Jackson Laboratory	RRID: IMSR_JAX:008471
Zebrafish: Tg(Shha:GFP)t10: t10Tg	Neumann and Nuesslein-Volhard3	ZFIN: ZDB-GENO-060207–1
Arabidopsis: 35S::PIF4-YFP, BZR1-CFP	Wang et al.4	N/A
Arabidopsis: JYB1021.2: pS24(AT5G58010)::cS24:GFP(-G):NOS #1	NASC	NASC ID: N70450
Oligonucleotides
siRNA targeting sequence: PIP5K I alpha #1: ACACAGUACUCAGUUGAUA	This paper	N/A
Primers for XX, see Table SX	This paper	N/A
Primer: GFP/YFP/CFP Forward: GCACGACTTCTTCAAGTCCGCCATGCC	This paper	N/A
Morpholino: MO-pax2a GGTCTGCTTTGCAGTGAATATCCAT	Gene Tools	ZFIN: ZDB-MRPHLNO-061106–5
ACTB (hs01060665_g1)	Life Technologies	Cat#4331182
RNA sequence: hnRNPA1_ligand: UAGGGACUUAGGGUUCUCUCUAGGGACUUAGGGUUCUCUCUAGGGA	This paper	N/A
Recombinant DNA
pLVX-Tight-Puro (TetOn)	Clonetech	Cat#632162
Plasmid: GFP-Nito	This paper	N/A
cDNA GH111110	Drosophila Genomics Resource Center	DGRC:5666; FlyBase:FBcl0130415
AAV2/1-hsyn-GCaMP6- WPRE	Chen et al.5	N/A
Mouse raptor: pLKO mouse shRNA 1 raptor	Thoreen et al.6	Addgene Plasmid #21339
Software and algorithms
ImageJ	Schneider et al.7	https://imagej.nih.gov/ij/
Bowtie2	Langmead and Salzberg8	http://bowtie-bio.sourceforge.net/bowtie2/index.shtml
Samtools	Li et al.9	http://samtools.sourceforge.net/
Weighted Maximal Information Component Analysis v0.9	Rau et al.10	https://github.com/ChristophRau/wMICA
ICS algorithm	This paper; Mendeley Data	http://dx.doi.org/10.17632/5hvpvspw82.1
Other
Sequence data, analyses, and resources related to the ultra-deep sequencing of the AML31 tumor, relapse, and matched normal	This paper	http://aml31.genome.wustl.edu
Resource website for the AML31 publication	This paper	https://github.com/chrisamiller/aml31SuppSite

PHYSICAL SCIENCE TABLE WITH EXAMPLES FOR AUTHOR REFERENCE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
QD605 streptavidin conjugated quantum dot	Thermo Fisher Scientific	Cat#Q10101MP
Platinum black	Sigma-Aldrich	Cat#205915
Sodium formate BioUltra, ≥99.0% (NT)	Sigma-Aldrich	Cat#71359
Chloramphenicol	Sigma-Aldrich	Cat#C0378
Carbon dioxide (13C, 99%) (<2% 18O)	Cambridge Isotope Laboratories	CLM-185–5
Poly(vinylidene fluoride-co-hexafluoropropylene)	Sigma-Aldrich	427179
PTFE Hydrophilic Membrane Filters, 0.22 mm, 90 mm	Scientificfilters.com/Tisch Scientific	SF13842
Critical commercial assays
Folic Acid (FA) ELISA kit	Alpha Diagnostic International	Cat# 0365–0B9
TMT10plex Isobaric Label Reagent Set	Thermo Fisher	A37725
Surface Plasmon Resonance CM5 kit	GE Healthcare	Cat#29104988
NanoBRET Target Engagement K-5 kit	Promega	Cat#N2500
Deposited data
B-RAF RBD (apo) structure	This paper	PDB: 5J17
Structure of compound 5	This paper; Cambridge Crystallographic Data Center	CCDC: 2016466
Code for constraints-based modeling and analysis of autotrophic E. coli	This paper	https://gitlab.com/elad.noor/sloppy/tree/master/rubisco
Software and algorithms
Gaussian09	Frish et al.1	https://gaussian.com
Python version 2.7	Python Software Foundation	https://www.python.org
ChemDraw Professional 18.0	PerkinElmer	https://www.perkinelmer.com/category/chemdraw
Weighted Maximal Information Component Analysis v0.9	Rau et al.2	https://github.com/ChristophRau/wMICA
Other
DASGIP MX4/4 Gas Mixing Module for 4 Vessels with a Mass Flow Controller	Eppendorf	Cat#76DGMX44
Agilent 1200 series HPLC	Agilent Technologies	https://www.agilent.com/en/products/liquid-chromatography
PHI Quantera II XPS	ULVAC-PHI, Inc.	https://www.ulvac-phi.com/en/products/xps/phi-quantera-ii/

Highlights.

• A chromosome-level genome assembly of the diurnal rodent R. pumilio was generated

• R. pumilio shows a realigned circadian organisation compared to nocturnal rodents

• There has been a relaxation of purifying selection in phototransduction genes

• R. pumilio and nocturnal rodents have differences in dim light visual responses

Inclusion and diversity:

We support inclusive, diverse, and equitable conduct of research.