Back

Gene expression in African Americans and Latinos reveals ancestry-specific patterns of genetic architecture

Mak, A. C. Y.; Kachuri, L.; Hu, D.; Eng, C.; Huntsman, S.; Elhawary, J. R.; Gupta, N.; Gabriel, S.; Xiao, S.; Gui, H.; Keys, K. L.; Oni-Orisan, A.; Oh, S. S.; Seibold, M. A.; Rodriguez-Santana, J. R.; LeNoir, M. A.; Williams, L. K.; Borrell, L. N.; Gignoux, C. R.; Zaitlen, N. A.; Burchard, E. G.; Ziv, E.

2021-08-19 genetics
10.1101/2021.08.19.456901 bioRxiv
Show abstract

We analyzed whole genome and RNA sequencing data from 2,733 African American and Hispanic/Latino children to explore ancestry- and heterozygosity-related differences in the genetic architecture of whole blood gene expression. We found that heritability of gene expression significantly increases with greater proportion of African genetic ancestry and decreases with higher levels of Indigenous American ancestry, consistent with a relationship between heterozygosity and genetic variance. Among heritable protein-coding genes, the prevalence of statistically significant ancestry-specific expression quantitative trait loci (anc-eQTLs) was 30% in African ancestry and 8% for Indigenous American ancestry segments. Most of the anc-eQTLs (89%) were driven by population differences in allele frequency, demonstrating the importance of measuring gene expression across multiple populations. Transcriptome-wide association analyses of multi-ancestry summary statistics for 28 traits identified 79% more gene-trait pairs using models trained in our admixed population than models trained in GTEx. Our study highlights the importance of large and ancestrally diverse genomic studies for enabling new discoveries of complex trait architecture and reducing disparities.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.