Organoid eQTLs can help interpret human disease genetics. (A) Example topic-interacting eQTLs. ABCA1 expression is correlated with the inferred cell environment, as defined by linear combinations of topics (Bonferroni-corrected P-value 0.00277), and this effect is largely driven by cortical hem and glial progenitor topic 15 in a genotype-dependent manner. The WDR45B gene-by-context eQTL effect (Bonferroni-corrected P-value 0.00184) is largely explained by hypoxia-associated topic 7. Each point corresponds to a single pseudocell used for CellRegMap topic interaction QTL mapping. (B) Fraction of eGenes in each cell type which are the nearest gene to a genome-wide significant GWAS finding among 402 brain-related traits. (C) Example of a cell type– and treatment-specific regulatory association matching a significant GWAS variant (lfsr 0.54, 0.04, and 0.66, with posterior estimated effect sizes of −0.11, −0.55, and −0.02, for control, hypoxia, and hyperoxia conditions, respectively). (D) Number of eGenes in each cell type and discovery condition for which rare loss-of-function alleles have been associated with disease (see Methods).
