(Porig = 8.91 × 10−6; Pcond = 2.0 × 10−2) and DPW (Porig = 4.90 × 10−12; Pcond = 7.0 × 10−4) even after adding rs10838753 as a covariate. Rs56030824 overlapped the promoter marks (H3K27ac, H3K4me3) for SPI1 specifically in microglia (Fig. 5A). This SNP also alters the binding site regulatory motif for RXRA, a transcription factor which is involved in the promotion of myelin debris phagocytosis and remyelination by macrophages24. Since SPI1 is expressed in myeloid lineage cells, its mRNA expression in the bulk brain was too low to perform differential expression or integration analysis. Therefore, we chose eQTLs from a large sample of peripheral blood monocytes to examine if rs56030824 is associated with the expression of SPI1 in these cells. The effect sizes of eQTLs at 11p11.2 locus were linearly correlated with effect sizes from the DPW GWAS at this locus (Fig. 5B, C). In fact, rs56030824 had the strongest effect size for SPI1 expression and DPW in the common variant category (Fig. 5C). These observations together established rs56030824 as a stronger candidate to be considered as a causal variant and SPI1 as a potential candidate gene associated with AUD and DPW.
