Supplementary Figure 4: Combinatorial perturbation of SNAP91, TSNARE1, CLCN3 and FURIN.

A. Mean difference (MA) plots of differential expression in SNAP91, TSNARE1, CLCN3 CRISPRa, FURIN RNAi, the additive model of these perturbations, the combinatorial perturbation and the resulting synergistic effect in NGN2-excitatory neurons, with significantly differentially expressed genes highlighted (empirical Bayes method + multiple testing correction, FDR <5%). B. Heat map detailing the contrasts applied to the SZ-eQTL gene modified samples and controls to produce differential expression models as in (A). C. Expression variance is partitioned into fractions attributable to each experimental variable. Violin plots show the percentage of variance explained by each variable over all genes (n=19,469) and samples (n=34), with violin width indicating density estimate. Overlaid boxplots show first to third quartile and the median. Whiskers indicate largest/smallest observation (no larger/smaller than hinge +/- 1.5 * inter-quartile range). D. Pie chart showing percentages of genes that fall into different synergistic differential expression categories (n=19,469 genes total). E. Hierarchical clustering of the differential expression log2 (fold changes) of categories in (D), in the additive model vs. the combinatorial perturbation. CRMP1 and DLX1, as seen in Fig. 5b, are part of the ‘less up’ and ‘same up’ categories, respectively. F. Over-representation analysis (ORA)-derived significant module enrichment (FDR<5%) for ranked genes in the ‘more up’ and ‘more down’ synergistic effect categories.