Abstract
Gene regulatory innovation underlies many phenotypic transitions. Transposable elements are an established mechanism for creating families of cis-acting elements with shared sequence features and the potential to establish co-regulatory networks. To understand additional mechanisms by which co-regulatory networks form, we define families of noncoding elements based on sequence similarity and cell type-specific activity. We apply this analysis framework to the human telomere-to-telomere genome assembly and embryonic stem cell chromatin accessibility data. We identify segmental duplications as the major mechanism establishing these families, creating over one thousand networks of elements with open chromatin in embryonic stem cells. We functionally validate a subset of these networks as families of regulatory elements with STARR-seq and identify their target genes with CRISPRi in embryonic stem cells. Following segmental duplication, we find that regulatory elements at times maintain their relationship to target genes, and at times rewire to form novel connections. During this rewiring, we observe proximal-acting elements gaining the ability to regulate distally-located genes and observe transcriptional enhancers rewiring to regulate genes present at the locus outside the segmental duplication. Many of these rewiring events are human specific. Finally, we find that segmental duplications have made outsized contributions to expanding regulatory element families functioning in immune cell types and specific brain regions, including the posterior cingulate gyrus. We speculate that placing regulatory elements in new genomic contexts primes regulatory elements for neofunctionalization, and that regulatory rewiring after segmental duplication was a common mechanism underlying gene regulatory change during human evolution.
Full text
2,044 characters
· extracted from
oa-doi-fallback
· click to expand
Abstract
Gene regulatory innovation underlies many phenotypic transitions. Transposable elements are an established mechanism for creating families of cis-acting elements with shared sequence features and the potential to establish co-regulatory networks. To understand additional mechanisms by which co-regulatory networks form, we define families of noncoding elements based on sequence similarity and cell type-specific activity. We apply this analysis framework to the human telomere-to-telomere genome assembly and embryonic stem cell chromatin accessibility data. We identify segmental duplications as the major mechanism establishing these families, creating over one thousand networks of elements with open chromatin in embryonic stem cells. We functionally validate a subset of these networks as families of regulatory elements with STARR-seq and identify their target genes with CRISPRi in embryonic stem cells. Following segmental duplication, we find that regulatory elements at times maintain their relationship to target genes, and at times rewire to form novel connections. During this rewiring, we observe proximal-acting elements gaining the ability to regulate distally-located genes and observe transcriptional enhancers rewiring to regulate genes present at the locus outside the segmental duplication. Many of these rewiring events are human specific. Finally, we find that segmental duplications have made outsized contributions to expanding regulatory element families functioning in immune cell types and specific brain regions, including the posterior cingulate gyrus. We speculate that placing regulatory elements in new genomic contexts primes regulatory elements for neofunctionalization, and that regulatory rewiring after segmental duplication was a common mechanism underlying gene regulatory change during human evolution.
Competing Interest Statement
C.B.L. owns stock in Alphabet and has a family member and friends who are employees of Alphabet subsidiaries. The other authors report no conflicts of interest.
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.