Human accelerated regions drive unique expression patterns during embryonic development

Katie Pollard
Gladstone Institutes, UCSF

(April 22, 2013 3:00 PM - 3:50 PM)

Human accelerated regions drive unique expression patterns during embryonic development

Abstract

The dramatic diversity of form and function found between closely related species is likely driven by changes to non-coding DNA that modify the complex patterns of gene expression observed throughout development. To pinpoint regions of the human genome that evolved rapidly since divergence from the chimp-human ancestor, we developed a statistical phylogenetic method for detecting lineage-specific changes in the rate or pattern of nucleotide substitutions. We analyzed vertebrate whole-genome multiple sequence alignments and found 721 Human Accelerated Regions (HARs). The vast majority of HARs are located in unannotated non-coding regions of the human genome. However, they are enriched nearby transcription factors and developmental genes, and many have epigenetic marks and transcription factor binding sites suggestive of enhancer function. To test this hypothesis we trained a multi-kernel support vector machine using experimentally validated developmental enhancers and diverse feature data (e.g., k-mers, transcription factor (TF) binding sites, cell type specific histone modifications, chromatin state). We predicted that ~2% of the human genome and over 200 HARs function as enhancers in different embryonic tissues. To explore whether human mutations in HARs alter their function, we developed a novel measure of regulatory sequence divergence based on cumulative loss and gain of predicted TF binding sites and showed that it identifies enhancers whose mutations affect activity in vivo. We used regulatory divergence in combination with expression patterns and functions of nearby genes to predict which candidate HAR enhancers are most likely to affect human-specific developmental gene regulation. We tested 15 of these predictions with transient transgenic mouse enhancer assays that compare activity of ancestral and derived HAR sequences. We found many novel developmental enhancers, several of which have human-specific activity.