Genome-wide identification and characterization of QTLs for transcriptional noise in human midbrain cells

While cell-to-cell variation in gene expression, also known as "transcriptional noise," plays various biological roles, its regulation by genetic variants remains elusive. To address this, we analyzed single-cell RNA sequencing (scRNA-seq) data of induced pluripotent stem cell-derived midbrain cells from 155 individuals together with their genotypes and identified significant quantitative trait loci for transcriptional noise (tnQTLs) genome wide. Among these, tnQTLs without significant effects on expression abundance (termed tn>eQTLs) exhibited characteristics such as drastic alterations under oxidative stress. Analyses using genome-wide association study (GWAS) summary statistics identified enrichment of schizophrenia association signals in tn>eQTLs, as well as putative causal effects of transcriptional noise dysregulation in specific genes. We also analyzed brain scRNA-seq data for schizophrenia and found marked disease-associated transcriptional noise alterations in superficial- and deep-layer excitatory neurons. Overall, this study provides a resource for tnQTLs and insights into the mechanistic basis of transcriptional noise regulation and its relevance to human traits.