Diverse molecular mechanisms contribute to differential expression of human duplicated genes

Emerging evidence links genes within human-specific segmental duplications (HSDs) to traits and diseases unique to our species. Strikingly, despite being nearly identical by sequence (>98.5%), paralogous HSD genes are differentially expressed across human cell and tissue types, though the underlying mechanisms have not been examined. Comparing cross-tissue mRNA levels between humans and chimpanzees, we determined that broadly, HSD paralogs (75 genes from 30 families) display expression patterns consistent with pseudo- or neofunctionalization. In general, the ancestral paralog exhibited greatest expression conservation with the chimpanzee ortholog, though exceptions suggest duplicate paralogs that may retain or supplant ancestral functions. To understand mechanisms underlying this observed regulatory divergence, we reanalyzed data from human lymphoblastoid cell lines (LCLs) (n=445), showing that ~75% of derived HSD paralogs exhibit significant differential expression and a greater than two-fold difference from their ancestral counterpart. To identify active cis-regulatory elements (CREs) in HSDs, we reanalyzed ENCODE data to recover hundreds of candidate CREs in these regions. Further, we generated ChIP-seq data for active chromatin features in an LCL using longer Illumina reads to better distinguish peaks in paralogous regions. Some of these duplicated CREs are sufficient to drive differential reporter activity, suggesting they may contribute to divergent cis-regulation of paralogs. This work provides evidence that cis-regulatory divergence contributes to novel expression patterns of recent gene duplicates in humans.