Distribution and intensity of constraint in mammalian genomic sequence

Abstract

Comparisons of orthologous genomic DNA sequences can be used to characterize regions that have been subject to purifying selection and are enriched for functional elements. We here present the results of such an analysis on an alignment of sequences from 29 mammalian species. The alignment captures ∼3.9 neutral substitutions per site and spans ∼1.9 Mbp of the human genome. We identify constrained elements from 3 bp to over 1 kbp in length, covering ∼5.5% of the human locus. Our estimate for the total amount of nonexonic constraint experienced by this locus is roughly twice that for exonic constraint. Constrained elements tend to cluster, and we identify large constrained regions that correspond well with known functional elements. While constraint density inversely correlates with mobile element density, we also show the presence of unambiguously constrained elements overlapping mammalian ancestral repeats. In addition, we describe a number of elements in this region that have undergone intense purifying selection throughout mammalian evolution, and we show that these important elements are more numerous than previously thought. These results were obtained with Genomic Evolutionary Rate Profiling (GERP), a statistically rigorous and biologically transparent framework for constrained element identification. GERP identifies regions at high resolution that exhibit nucleotide substitution deficits, and measures these deficits as "rejected substitutions." Rejected substitutions reflect the intensity of past purifying selection and are used to rank and characterize constrained elements. We anticipate that GERP and the types of analyses it facilitates will provide further insights and improved annotation for the human genome as mammalian genome sequence data become richer.

Optimized filtering reduces the error rate in detecting genomic variants by short-read sequencing

“…The first SNV located on chromosome 4 (position 189,411,955) is heterozygous in the healthy twin. This variant is located in a genomic evolutionary rate profiling (GERP) constraint element of 135 bp with the closest gene >100 kb away 31 . Intriguingly, the second SNV turned out to be a mosaic variant with a higher frequency of the variant allele in the schizophrenic twin.…”

Accurate molecular diagnosis of phenylketonuria and tetrahydrobiopterin-deficient hyperphenylalaninemias using high-throughput targeted sequencing

“…Only SNVs and InDels of up to 30 bp and found within 150 bp of the ends of the enriched targets were considered for subsequent analysis. Functional annotation of high-quality variants was performed using Annovar, 11 providing a comparison of the predicted variants to the National Center for Biotechnology Information (NCBI) SNP Database build 132 (dbSNP132), the March 2010 pilot release of the 1000 Genomes project (1000G; www.1000genomes.org), conservation around variants based on phastCons, 12 segmental duplication filter, gene annotation (exon/intron/UTR), amino-acid substitutions and splice variants based on UCSC Genome Browser 13 tracks, as well as multiple estimates of the impact of amino-acid substitution on the structure and function of proteins (tools: Sift, 14 Polyphen2, 15 PhyloP, 16 and MutationTaster 17 ). The reference sequences used for the four genes targeted in this study were NM_000277 (PAH), NM_000320 (QDPR), NM_000161 (GCH1), and NM_000317 (PTS).…”

“…4. Variants were ranked based on evolutionary conservation and potential deleteriousness of the affected nucleotide using Sift, 14 Polyphen2, 15 PhyloP, 16 and MutationTaster. 17 All newly identified variants identified in this study have been submitted to PAHdb (http://www.pahdb.mcgill.ca/), which is the reference database for hyperphenylalaninemia mutations.…”

New Methods for Detecting Lineage-Specific Selection

“…In the comparative genomics community, much attention has focused on two problems in particular: (1) identifying (especially noncoding) sequences that are unusually conserved across species, and thus are likely to be subject to negative selection (e.g., [3][4][5][6]); and (2) identifying protein-coding genes that show unusually high d N /d S ratios, and thus might be subject to positive selection (e.g., [7][8][9][10][11][12]). Methods focused on problem (1) generally have made the assumption (explicitly or implicitly) that selectional pressures are the same across all branches of a phylogeny-i.e., that each candidate sequence is under selection in all species or not under selection in any species.…”