Abstract
Quantification of gene expression in cells or tissue can inform on etiology of disease. Complementing these approaches, we propose to estimate subject- and cell-type-specific (CTS) gene expression from tissue using an empirical Bayes method that borrows information across multiple measurements of the same tissue per subject. Analyzing multiple brain regions from the Genotype-Tissue Expression project (GTEx) reveals a subset of expression quantitative trait loci specific to neurons, others specific to astrocytes, and others active across all cell types. In another example, CTS expression of the BrainSpan atlas, which profiles expression patterns of the developing human brain, demonstrates potential insights into processes associated with neurodevelopmental disorders. Our analyses reveal clear CTS co-expression networks that, when combined with genetic findings in autism spectrum disorder (ASD), identify a cluster of co-expressed ASD-associated genes and implicate immature neurons in ASD risk.