Supplementary Figure 6: CRISPR validation of the BLK locus.

(a,b), Scatterplots of log₁₀ Bayes factors (BF) for SLE (Nat. Genet. 47, 1457–1464, 2015) and RA (Nature 506, 376–381, 2014) against the log₁₀ Bayes factor of the eQTL for the BLK gene. The color of each point represents the LD index/r² between a variant in the cis region and the index SNP (indicated by a purple point). (c,d), Regional plots of disease associations (based on log₁₀ Bayes factor) for SLE and RA. The color of each point is the r² value between a variant in the cis region and the index SNP (indicated by a purple point). (e–g), ATAC-seq coverage plots around the putative causal variants. Coverage is stratified by the three genotypes of the causal variants indicated in each panel. The sequence logo is drawn based on the known transcription factor binding motif in CISBP (Cell 158, 1431–1443, 2014). We selected the motif that gives the highest binding affinity difference between the reference and alternative alleles and the orientation of the affinity was also the same as the chromatin openness. (h), Aggregated ATAC-seq coverage around the BLK locus for the two deletion lines (green, D1; orange, D2) superimposed on the parental line (R; navy). (i), Average chromatin accessibility in terms of FPKM at each peak (square dots) segregated by the three different cell lines (navy, parental line; green, D1 heterozygous line; orange, D2 heterozygous line). The peaks surrounded by red rectangles showed a consistent pattern to the caQTL results. (j), Expression levels of the FAM167A gene for the three different cell lines with two replications (n = 6 replicates in total; see section 5.7 in the Supplementary Note for details of statistical hypothesis testing).