Figure 2.

Isoform-Specific Perturb-Seq enables large-scale transcript-specific knockdown (A) A schematic of the pipeline of Promoter-Specific Perturb-Seq from promoter identification (left), generation of dual guides for respective P1 and P2 promoters and creation of MCF-7 stable cell line (centre) to 10× single-cell sequencing (right). (B) Plot of genomic and transcribed genome distances between P1 and P2 promoters for the candidate genes targeted in the screen. The transcribed genome distance represents the linear distance between the transcription start sites of P1 and P2 along the exon-only canonical Ensembl transcript annotation. This distance reflects the effective transcribed span separating the promoters within the reference transcript structure. (C) (Top) A genome browser track of ESR1 with P1 (green) and P2 (pink) promoters annotated. Top track displaying the targeted primers. Only transcripts with >5 TPM from bulk RNA-seq shown. (Bottom) Quantitative RT-qPCR results across sgRNA P1 and P2 from primers targeted to a region common to all estrogen receptor (ESR1) transcripts (left) and primers specific to P1 transcripts (right) pooled from two technical and two biological replicates. Blue line represents the NTC sgRNA relative expression. Relative Expression (ΔΔCt) relative to GADPH, a housekeeping gene (from left to right P = 0.0108, P = 0.221, P = 0.934 and P = 0.591, one-sample t-test). The blue dotted line represents the NTC. Significance codes P >.05 signified by n.s. (not significant), P < 0.05 * (D) Violin plots of percentage knockdown of targeted promoters associated transcripts relative to NTCs for sgRNAs targeting P1 promoters relative to transcripts produced from targeted and non-targeted promoters (left, n = 35, NTC TPM > 0.5, ***P < 6.3 × 10 − 5, two-sided Welch’s t-test) and P2 promoters (right, n = 35, NTC TPM > 0.5, *P < 5 × 10 − 2, two-sided Welch’s t-test) (%KD = 100 × (sum TPM NTC​/sum TPM KD​​ − 1). (E) Violin plots showing a statistical quantification of transcriptome distances between single cells with successful P1- to P2-sgRNAs knockdowns (n = 31). TOP plot shows adjusted P-values (Monte–Carlo permutation test with Holm–Sidak multiple test correction). Bottom plot shows putative changes to mRNA or protein transcript isoform defined by if canonical ATG resides between P1 and P2. (F) UMAP plots for BRCA1-associated C-terminal helicase 1 (BRIP1), Myb-binding protein 1 (MYBB1A), Estrogen Receptor 1 (ESR1) and proteosome 26S subunit (PSMC5) showing different clustering of single cells knockdown by P1 and P2 promoters. (G) A dot plot displaying the differential gene expression (DEG) analysis for P1 and P2 promoters against NTCs. Numbers of P1 and P2 differential genes for each gene are shown with a central number displaying several overlapping significantly differentially expressed genes (pval < 0.05, log FC > 0.5, genes exp > 10% of cells, yellow lettering signifies sig. e-distance) between P1 and P2.