JSON
Naidoo T, Sjödin P, Schlebusch C, Jakobsson M
BMC Genomics 19 (1) 95 [2018-01-26; online 2018-01-26]
With only 2 % of the human genome consisting of protein coding genes, functionality across the rest of the genome has been the subject of much debate. This has gained further impetus in recent years due to a rapidly growing catalogue of genomic elements, based primarily on biochemical signatures (e.g. the ENCODE project). While the assessment of functionality is a complex task, the presence of selection acting on a genomic region is a strong indicator of importance. In this study, we apply population genetic methods to investigate signals overlaying several classes of regulatory elements. We disentangle signals of purifying selection acting directly on regulatory elements from the confounding factors of demography and purifying selection linked to e.g. nearby protein coding regions. We confirm the importance of regulatory regions proximal to coding sequence, while also finding differential levels of selection at distal regions. We note differences in purifying selection among transcription factor families. Signals of constraint at some genomic classes were also strongly dependent on their physical location relative to coding sequence. In addition, levels of selection efficacy across genomic classes differed between African and non-African populations. In order to assign a valid signal of selection to a particular class of genomic sequence, we show that it is crucial to isolate the signal by accounting for the effects of demography and linked-purifying selection. Our study highlights the intricate interplay of factors affecting signals of selection on functional elements.
PubMed 29373957
DOI 10.1186/s12864-017-4422-y
Crossref 10.1186/s12864-017-4422-y
pii: 10.1186/s12864-017-4422-y
pmc: PMC5787233