Fig. 9: Allele specific binding events and transcriptional activity.

a For skin disease associated SNVs, there is a positive correlation in JUN CUT&RUN reads for risk vs protective alleles with transcriptional activity from the same alleles. b ASB fold changes may be greater on daSNVs. ASBs were categorized by whether they were daSNVs in MPRA, and by genomic location. The P value is from a two-sided t-test. Outlier values were clipped to 2 for visualization. Whiskers, max value within 1.5 interquartile ranges. Center line, median. c ASB and transcriptional changes are not significantly decreased on risk alleles. d Overlap of putative ASB with skin disease associated MPRA shows potential sites of ASB-driven changes in transcriptional activity at homeostasis genes. Allele freq., frequency of risk allele (crossed white box if the RSID was not found). daSNV, FDR < 0.01. e rs4704864 is an eQTL for the monogenic skin disease gene NIPAL4 and breaks a JUN motif. f The risk allele results in reduced transcriptional activity in MPRA (FDR 0.0009), g JUN binding in CUT&RUN (FDR 0.0009). h NIPAL4 RNA expression by qPCR after genomic edits. Cas9-AAV-based homology directed repair gene editing (HDR). THG1L is a nearby control gene. Whiskers, 95% CI. i–k The risk allele of rs10217259 results in increased transcriptional activity (FDR 0.004) and increased JUN binding (FDR 0.18). rs10217259 is looped to the KLF4 promoter and is associated with psoriasis. l–n rs1264326 is associated with rosacea and located adjacent to the DDR1 gene promoter; the risk allele exhibited more transcriptional variability and TFAP2A binding (FDR 0.02), although the risk allele was not significantly increased or decreased in MPRA activity. o Summary of screening, bulk RNA-seq, and MPRA results. p Provisional model of one potential mechanism of transcriptional control and disease-linked noncoding variant function in epidermal cells with the following features. Homeostatic TFs bind (and exhibit ASBs) near differentiation and monogenic disease-associated genes, including KLF4, TP63 and NIPAL4. The AP-1/2 and SP/KLF families may help convert SNVs to altered binding and transcription. The promoter region is buffered against binding changes, especially when many TFs are bound. Risk alleles are not greatly biased to loss-of-binding or activity. AP-1 ASBs predict transcriptional changes. Noncoding SNVs act through altered homeostatic TF binding to alter differentiation and disease risk.