Fig. 5: PRDM9-dependent H3K4me3 marks cholesterol biosynthesis genes.

a, b Heatmap and Gene Ontology of top 50 down-regulated DEGs (RNA sequencing of n = 3 biological replicates) in RKI1 cells treated with CMPD1 (10 µM, 3 days) ± MRK-740 (3 µM). P-adjusted value (Benjamin-Hochberg correction) from Fisher’s exact test. c Waterfall plot of differentially expressed genes (RNA sequencing of n = 3 biological replicates) in CMPD1 (10 µM, 3 days) treated RKI1 cells transduced with NTC sgRNA (NTC) or PRDM9 sgRNA (PRDM9 KO). d Genome wide H3K4me3 signal intensity (RPGC) (± 3 kb from transcription start/end site) in RKI1 cells treated with CMPD1 (10 µM, 3 days) ± MRK-740 (3 µM). e 2D kernel density plot showing the relationship between changes in H3K4me3 peak width and mRNA expression in RKI1 cells treated with CMPD1 (10 µM, 3 days) ± MRK-740 (3 µM). Each data point corresponds to a gene that was detected by RNA sequencing and contained a H3K4me3 peak with > 1000 bp in CMPD1-treated cells. The colour bar reflects the density. Pearson’s product-moment correlation was used to calculate the correlation coefficient (R) and statistical significance (P). f H3K4me3 signal intensity (RPGC) for down-regulated genes (listed in Fig. 5a; ± 3 kb from transcription start/end site) in RKI1 cells treated with CMPD1 (10 µM, 3 days) ± MRK-740 (3 µM). g Scatter plot of top 39 down-DEGs listed in Fig. 5a and containing a H3K4me3 peak > 1000 bp in CMPD1-treated cells. The plot shows correlation between changes in H3K4me3 peak width and mRNA expression in RKI1 cells treated with CMPD1 (10 µM, 3 days) ± MRK-740 (3 µM). Pearson’s product-moment correlation was used to calculate the correlation coefficient (R) and statistical significance (P). h Venn diagram of functional genomic PRDM9 annotated regions (Homo Sapiens PRDM9 matrix; Supplementary Data 2) and down-regulated genes listed in Fig. 5a. i H3K4me3 genome tracks for cholesterol biosynthesis genes in RKI1 cells treated with CMPD1 (10 µM, 3 days) ± MRK-740 (3 µM).