CoChat
Sign Up
Cited by 578Evolution and Genetic DynamicsGenomics and Phylogenetic StudiesGenetic diversity and population structure

The DOI matches its Crossref record; 95 of the 97 references it cites were found in a registry or an index.

Patterns of Positive Selection in Six Mammalian Genomes

Open access
Kosiol, Carolin, Vinař, Tomáš, da Fonseca, Rute R., Hubisz, Melissa J., Bustamante, Carlos D., Nielsen, Rasmus, Siepel, Adam·2008·PLoS Genetics
DOI 10.1371/journal.pgen.1000144PMC PMC2483296
Open this paper in CoChat — free accountOpen PDF

Abstract

Genome-wide scans for positively selected genes (PSGs) in mammals have provided insight into the dynamics of genome evolution, the genetic basis of differences between species, and the functions of individual genes. However, previous scans have been limited in power and accuracy owing to small numbers of available genomes. Here we present the most comprehensive examination of mammalian PSGs to date, using the six high-coverage genome assemblies now available for eutherian mammals. The increased phylogenetic depth of this dataset results in substantially improved statistical power, and permits several new lineage- and clade-specific tests to be applied. Of approximately 16,500 human genes with high-confidence orthologs in at least two other species, 400 genes showed significant evidence of positive selection (FDR<0.05), according to a standard likelihood ratio test. An additional 144 genes showed evidence of positive selection on particular lineages or clades. As in previous studies, the identified PSGs were enriched for roles in defense/immunity, chemosensory perception, and reproduction, but enrichments were also evident for more specific functions, such as complement-mediated immunity and taste perception. Several pathways were strongly enriched for PSGs, suggesting possible co-evolution of interacting genes. A novel Bayesian analysis of the possible "selection histories" of each gene indicated that most PSGs have switched multiple times between positive selection and nonselection, suggesting that positive selection is often episodic. A detailed analysis of Affymetrix exon array data indicated that PSGs are expressed at significantly lower levels, and in a more tissue-specific manner, than non-PSGs. Genes that are specifically expressed in the spleen, testes, liver, and breast are significantly enriched for PSGs, but no evidence was found for an enrichment for PSGs among brain-specific genes. This study provides additional evidence for widespread positive selection in mammalian evolution and new genome-wide insights into the functional implications of positive selection.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1371/journal.pgen.1000144

Cites

95 of 95 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 The LRTs used to detect positive selection in the six mammalian genomes. (A–I) Panel A shows the test for selection on any branch of the phylogeny, and panels B–I show the lineage- and clade-specific tests, with branches under positive selection highlighted. The numbers below each subfigure represent the number of positively selected genes identified by each LRT (FDR<0.05) and the total number of ortholog sets tested. In (A), branch lengths are drawn proportional to their estimates in substitutions per site, and each branch is labeled with the corresponding estimate of ω. All tests are based on an unrooted phylogeny; the trees are rooted for display purposes only. Nominal P-value thresholds for FDR<0.05 were: (A) 1.1×10−3, (B) 9.1×10−5, (C) 7.7×10−5, (D) 2.9×10−4, (E) 2.8×10−4, (F) 2.5×10−5, (G) 5.4×10−5, (H) 1.8×10−5, (I) 5.9×10−5.
Figure 1 The LRTs used to detect positive selection in the six mammalian genomes. (A–I) Panel A shows the test for selection on any branch of the phylogeny, and panels B–I show the lineage- and clade-specific tests, with branches under positive selection highlighted. The numbers below each subfigure represent the number of positively selected genes identified by each LRT (FDR<0.05) and the total number of ortholog sets tested. In (A), branch lengths are drawn proportional to their estimates in substitutions per site, and each branch is labeled with the corresponding estimate of ω. All tests are based on an unrooted phylogeny; the trees are rooted for display purposes only. Nominal P-value thresholds for FDR<0.05 were: (A) 1.1×10−3, (B) 9.1×10−5, (C) 7.7×10−5, (D) 2.9×10−4, (E) 2.8×10−4, (F) 2.5×10−5, (G) 5.4×10−5, (H) 1.8×10−5, (I) 5.9×10−5.CC BY · PubMed Central
Figure 2 Hierarchical clustering of 27 over-represented GO categories identified by the Mann-Whitney U test (“biological process” group only), based on the genes assigned to each category. This dendrogram is derived from a dissimilarity matrix defined such that any two GO categories, X and Y, have dissimilarity 0 when all genes assigned to X are also assigned to Y (or vice-versa), and dissimilarity 1 when the sets of genes assigned to X and Y do not overlap. Specifically, X and Y have dissimilarity , where denotes the (nonempty) set of genes assigned to GO category C. Thus, GO categories associated with similar sets of genes group together in the dendrogram, even if these categories are not closely related in the GO hierarchy (such as “cytolysis” and “single fertilization”). Full names of abbreviated categories (*) are “humoral immune response mediated by circulating immunoglobulin,” “activation of plasma proteins during acute inflammatory response,” and “adaptive immune response based on somatic recombination of immune receptors built from immunoglobulin superfamily domains.” (Dendrogram produced by the hclust function in R with method = “average”.)
Figure 2 Hierarchical clustering of 27 over-represented GO categories identified by the Mann-Whitney U test (“biological process” group only), based on the genes assigned to each category. This dendrogram is derived from a dissimilarity matrix defined such that any two GO categories, X and Y, have dissimilarity 0 when all genes assigned to X are also assigned to Y (or vice-versa), and dissimilarity 1 when the sets of genes assigned to X and Y do not overlap. Specifically, X and Y have dissimilarity , where denotes the (nonempty) set of genes assigned to GO category C. Thus, GO categories associated with similar sets of genes group together in the dendrogram, even if these categories are not closely related in the GO hierarchy (such as “cytolysis” and “single fertilization”). Full names of abbreviated categories (*) are “humoral immune response mediated by circulating immunoglobulin,” “activation of plasma proteins during acute inflammatory response,” and “adaptive immune response based on somatic recombination of immune receptors built from immunoglobulin superfamily domains.” (Dendrogram produced by the hclust function in R with method = “average”.)CC BY · PubMed Central
Figure 3 Structural analysis of the HAVCR1 gene. At top is a graph showing the domain structure of the gene and corresponding Bayes Empirical Bayes [83] posterior probabilities (PP) of positive selection, based on our six-species alignments, with sites predicted to be under positive selection (PP>0.95) in red. At bottom right is a structural diagram (based on the structure of the IgV domain of the mouse gene) showing the interaction between two receptors that have been implicated in the regulation of HAVCR1's immune function. It is thought that clustering of receptors within the same cell surface might facilitate phosphorylation of the cytoplasmic tail, and that interaction between receptors from different cells might be a mechanism for B–T cell adhesion [91]. Predicted residue 39 falls within the region of these receptors, very near residue 37, which directly interacts with the opposite receptor (according to the available mouse structure). In addition, predicted residues 54 and 56 are adjacent to the virus-binding surface (shown in pink), as defined by a polymorphism in macaque [91]. Interestingly, the residue that falls between them (55) appears to be critical for virus-binding at the homologous loop in the CEA coronavirus receptor [91]. Residue 75 in the IgV domain also shows evidence of positive selection (PP>0.90, shown in orange) but its function is unknown.
Figure 3 Structural analysis of the HAVCR1 gene. At top is a graph showing the domain structure of the gene and corresponding Bayes Empirical Bayes [83] posterior probabilities (PP) of positive selection, based on our six-species alignments, with sites predicted to be under positive selection (PP>0.95) in red. At bottom right is a structural diagram (based on the structure of the IgV domain of the mouse gene) showing the interaction between two receptors that have been implicated in the regulation of HAVCR1's immune function. It is thought that clustering of receptors within the same cell surface might facilitate phosphorylation of the cytoplasmic tail, and that interaction between receptors from different cells might be a mechanism for B–T cell adhesion [91]. Predicted residue 39 falls within the region of these receptors, very near residue 37, which directly interacts with the opposite receptor (according to the available mouse structure). In addition, predicted residues 54 and 56 are adjacent to the virus-binding surface (shown in pink), as defined by a polymorphism in macaque [91]. Interestingly, the residue that falls between them (55) appears to be critical for virus-binding at the homologous loop in the CEA coronavirus receptor [91]. Residue 75 in the IgV domain also shows evidence of positive selection (PP>0.90, shown in orange) but its function is unknown.CC BY · PubMed Central
Figure 4 Patterns of positive selection on the mammalian phylogeny. (A) Probabilities that each gene gains (green) or loses (red) positive selection on each branch, under the Bayesian switching model. Switching events are allowed to occur early (near ancestor) or late (near descendant) on internal branches, and early on external branches. The prior probability of selection at the root of the tree is shown in parentheses. (The primate-rodent ancestor is treated as the root for this analysis; see Text S1.) The values shown are posterior means. The full posterior distributions are summarized in Figure S2. (B) Expected numbers of genes under positive selection on each branch (blue) and under positive selection only on each branch (red), out of the 544 PSGs examined, with 95% credible intervals in parentheses. Branch thicknesses are proportional to numbers in blue. Similar estimates are also shown for genes under positive selection on all branches of the primate and rodent clades (blue), on only the branches of these clades (red), and on all branches of the tree (blue). All estimates are based on 10,000 iterations of the Gibbs sampler, excluding a 100 iteration burn-in period. On each iteration, all switching parameters and the selection histories for all genes were sampled (see Text S1).
Figure 4 Patterns of positive selection on the mammalian phylogeny. (A) Probabilities that each gene gains (green) or loses (red) positive selection on each branch, under the Bayesian switching model. Switching events are allowed to occur early (near ancestor) or late (near descendant) on internal branches, and early on external branches. The prior probability of selection at the root of the tree is shown in parentheses. (The primate-rodent ancestor is treated as the root for this analysis; see Text S1.) The values shown are posterior means. The full posterior distributions are summarized in Figure S2. (B) Expected numbers of genes under positive selection on each branch (blue) and under positive selection only on each branch (red), out of the 544 PSGs examined, with 95% credible intervals in parentheses. Branch thicknesses are proportional to numbers in blue. Similar estimates are also shown for genes under positive selection on all branches of the primate and rodent clades (blue), on only the branches of these clades (red), and on all branches of the tree (blue). All estimates are based on 10,000 iterations of the Gibbs sampler, excluding a 100 iteration burn-in period. On each iteration, all switching parameters and the selection histories for all genes were sampled (see Text S1).CC BY · PubMed Central
Figure 5 Distributions of expression levels in PSGs (red) and non-PSGs (blue) for three tissue types. (A–C) Distributions as estimated from Affymetrix Human Exon 1.0 ST Array data by the RMA algorithm [88]. The other eight tissue types showed similar differences between PSGs and non-PSGs (Figure S7). (D) Distribution of degree of tissue bias in expression levels for PSGs (red) and non-PSGs (blue), as measured by the statistic τ [61] (Methods). An alternative measure of tissue bias (γ) showed a similar pattern.
Figure 5 Distributions of expression levels in PSGs (red) and non-PSGs (blue) for three tissue types. (A–C) Distributions as estimated from Affymetrix Human Exon 1.0 ST Array data by the RMA algorithm [88]. The other eight tissue types showed similar differences between PSGs and non-PSGs (Figure S7). (D) Distribution of degree of tissue bias in expression levels for PSGs (red) and non-PSGs (blue), as measured by the statistic τ [61] (Methods). An alternative measure of tissue bias (γ) showed a similar pattern.CC BY · PubMed Central
Figure 6 Power of the LRT for selection on any branch of the phylogeny as a function of the nonsynonymous-synonymous rate ratio ω. Power is defined as the fraction of tests resulting in nominal P<0.05. (The effect of controlling for multiple comparisons is shown in Figure S3.) When ω≤1, these fractions are estimates of the false positive rate. Each data point is based on 1000 data sets simulated with evolver [84] under the assumption of a constant ω among lineages and among sites (model M0). All other parameters (including the transition-transversion ratio κ, the codon frequencies, and the branch lengths) were fixed at values estimated from the real data. Results are shown for short (200-codon) and long (500-codon) genes and three sets of species: hominids (human and chimpanzee), primates (human, chimpanzee, and macaque), and all six mammals. Details on the computation of P-values are given in Text S1. Note the logarithmic scale on the x-axis.
Figure 6 Power of the LRT for selection on any branch of the phylogeny as a function of the nonsynonymous-synonymous rate ratio ω. Power is defined as the fraction of tests resulting in nominal P<0.05. (The effect of controlling for multiple comparisons is shown in Figure S3.) When ω≤1, these fractions are estimates of the false positive rate. Each data point is based on 1000 data sets simulated with evolver [84] under the assumption of a constant ω among lineages and among sites (model M0). All other parameters (including the transition-transversion ratio κ, the codon frequencies, and the branch lengths) were fixed at values estimated from the real data. Results are shown for short (200-codon) and long (500-codon) genes and three sets of species: hominids (human and chimpanzee), primates (human, chimpanzee, and macaque), and all six mammals. Details on the computation of P-values are given in Text S1. Note the logarithmic scale on the x-axis.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Research works printed but not linked: 1

  1. JL Thorne, 2007, Population genetics without intraspecific data., Mol Biol Evol, 24, 1667, 10.1093/molbev/msm085 · DOI

Cited by

Sorted by citations

See all 578

Related

See all

Metadata

Authors
Kosiol, C., Vinař, T., da Fonseca, R. R., Hubisz, M. J., Bustamante, C. D., Nielsen, R., & Siepel, A.
Affiliations
  • Cornell University (US)
  • University of Chicago (US)
  • University of Copenhagen (DK)
Corresponding author
Siepel, Adam
ORCID
Venue
PLoS Genetics
Publisher
Public Library of Science (PLoS)
Published
2008-08-01
Volume
4
Issue
8
Pages
e1000144
ISSN
  • 1553-7390 (ISSN-L)
  • 1553-7404
DOI
10.1371/journal.pgen.1000144
PMC
PMC2483296
OpenAlex
W2154770881
Cited by
576
References
97
Citations by year
2026: 5 · 2025: 15 · 2024: 23 · 2023: 17 · 2022: 19 · 2021: 58 · 2020: 27 · 2019: 27 · 2018: 24 · 2017: 31 · 2016: 25 · 2015: 32 · 2014: 51 · 2013: 55 · 2012: 41
FWCI
23.75
Type
journal article
Language
en
Open access
Licence
Locations
  • PLoS Genetics (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Cold Spring Harbor Laboratory Institutional Repository (Cold Spring Harbor Laboratory) (acceptedVersion)
  • eScholarship (California Digital Library) (submittedVersion, cc-by)
  • Knowledge@UChicago (University of Chicago) (submittedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • eScholarship (California Digital Library) (submittedVersion)
  • Figshare (submittedVersion, cc-by)
  • Research at the University of Copenhagen (University of Copenhagen) (submittedVersion)
  • University of Chicago (cc-by)
Funders
  • National Science Foundation (0644111, 0516310)
  • Microsoft Research
  • National Human Genome Research Institute (R01HG003229)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Evolution and Genetic Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic diversity and population structure
Keywords
positive selection, mammalian evolution, placental mammals, orthologs, lineage-specific selection, immune-related genes, chemosensory perception, reproduction, taste perception, gene coevolution, episodic selection, gene expression, tissue-specific gene expression, spleen, testis, liver, breast
MeSH terms
Animals, Bayes Theorem, Pan troglodytes, Dogs, Humans, Macaca mulatta, Mammals, Phylogeny, Primates, Rodentia, Selection, Genetic, Gene Expression, Likelihood Functions, Sequence Alignment, Genome, Evolution, Molecular, Databases, Genetic, Mice, Rats

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review