Empirical profile mixture models for phylogenetic reconstruction

Le, Quang Si; Gascuel, Olivier; Lartillot, Nicolas

doi:10.1093/bioinformatics/btn445

Cited by 305 publications

(303 citation statements)

References 40 publications

Supporting

Mentioning

293

Contrasting

Order By: Relevance

“…However, because the 20, 40, or 60 profiles were obtained from a framework operating at the amino acid level only (i.e., that does not attempt to tease out mutational and selective effects), the Bayes factors associated with their use in the mutation-selection framework are likely to be lower than they might be with empirical mixtures obtained by incorporating the nucleotide (codon)-level data within the training methods described in ref. 12. Despite this weakness of the plug-in we attempt here, the natural log Bayes factors are ∼236, 255, and 269 for the MGMutSelC20, MG-MutSelC40, and MG-MutSelC60 models, respectively (computed against the MG model).…”

Section: Resultsmentioning

confidence: 93%

“…For these empirical versions, the mixture model is predefined, in terms of the number of components and in terms of the profiles of each component, which are fixed to those inferred in ref. 12; only the weights of the components are free parameters, endowed with a flat Dirichlet prior. Depending on the number of components, we refer to these models as MG-MutSelC20, MG-MutSelC40, and MG-MutSelC60.…”

Section: Resultsmentioning

confidence: 99%

“…The base prior probability vector is endowed with a Dirichlet hyperprior, and unidimensional parameters are endowed with exponential hyperpriors of mean 1. For the MG-MutSelC20, MG-MutSelC40, and MG-MutSelC60 models described in the main text, the only free parameters involved in the mixture are the weights of the components; these weights are implicitly integrated out via an allocation system, in a manner that is equivalent to having a flat Dirichlet prior on them (12). The priors on all other parameters are as described in ref.…”

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Mutation-selection models of coding sequence evolution with site-heterogeneous amino acid fitness profiles

Rodrigue

Piégay

Lartillot

2010

Proc. Natl. Acad. Sci. U.S.A.

Self Cite

156

194

View full text Add to dashboard Cite

Modeling the interplay between mutation and selection at the molecular level is key to evolutionary studies. To this end, codonbased evolutionary models have been proposed as pertinent means of studying long-range evolutionary patterns and are widely used. However, these approaches have not yet consolidated results from amino acid level phylogenetic studies showing that selection acting on proteins displays strong site-specific effects, which translate into heterogeneous amino acid propensities across the columns of alignments; related codon-level studies have instead focused on either modeling a single selective context for all codon columns, or a separate selective context for each codon column, with the former strategy deemed too simplistic and the latter deemed overparameterized. Here, we integrate recent developments in nonparametric statistical approaches to propose a probabilistic model that accounts for the heterogeneity of amino acid fitness profiles across the coding positions of a gene. We apply the model to a dozen real proteincoding gene alignments and find it to produce biologically plausible inferences, for instance, as pertaining to site-specific amino acid constraints, as well as distributions of scaled selection coefficients. In their account of mutational features as well as the heterogeneous regimes of selection at the amino acid level, the modeling approaches studied here can form a backdrop for several extensions, accounting for other selective features, for variable population size, or for subtleties of mutational features, all with parameterizations couched within population-genetic theory.codon substitution | Dirichlet process | phylogeny | selection coefficients F ollowing the seminal works of Muse and Gaut (1) and Goldman and Yang (2), most early applications of codon-based evolutionary models were focused on evaluations of selective effects operating at different positions along a gene or at different time points along the phylogeny (see refs. 3, 4 for reviews). Many of these approaches have modeled selective effects using a parameter representing the nonsynonymous/synonymous rate ratio. However, this may not be ideal, in particular because it amounts to ignoring differences between different pairs of possible amino acid replacements resulting from nonsynonymous point mutations. In recent years, questions regarding selective effects have diversified, such as in the work of Yang and Nielsen (5), who propose a test for selection on codon usage. This test is based on models that invoke a multidimensional specification of scaled selection coefficients, based on either 20 or 61 (under the universal genetic code) scaled fitness parameters-adding 19 or 60 degrees of freedom to the underlying codon substitution model-in contrast with the more conventional use of the single nonsynonymous/synonymous rate ratio parameter, viewing all nonsynonymous events as equivalent (e.g., see ref. 6). By assigning scaled fitness parameters to each of the 20 amino acids, or to the 61 sense codons, Yang and Nie...

show abstract

Section: Resultsmentioning

confidence: 93%

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Mutation-selection models of coding sequence evolution with site-heterogeneous amino acid fitness profiles

Rodrigue

Piégay

Lartillot

2010

Proc. Natl. Acad. Sci. U.S.A.

Self Cite

156

194

View full text Add to dashboard Cite

show abstract

“…However, a series of empirical observations suggest that finite mixtures are not sufficient. First, the empirical fit shows a substantial increase with the number of components even when this number is high [28,32], indicating that even relatively rich mixtures do not capture a sufficient fraction of the available variation. Forcing the number of categories to a relatively small number is possible.…”

Section: (C) Modelling Variation Across Sitesmentioning

confidence: 99%

Probabilistic models of eukaryotic evolution: time for integration

Lartillot¹

2015

Phil. Trans. R. Soc. B

Self Cite

View full text Add to dashboard Cite

One contribution of 17 to a theme issue 'Eukaryotic origins: progress and challenges'. In spite of substantial work and recent progress, a global and fully resolved picture of the macroevolutionary history of eukaryotes is still under construction. This concerns not only the phylogenetic relations among major groups, but also the general characteristics of the underlying macroevolutionary processes, including the patterns of gene family evolution associated with endosymbioses, as well as their impact on the sequence evolutionary process. All these questions raise formidable methodological challenges, calling for a more powerful statistical paradigm. In this direction, model-based probabilistic approaches have played an increasingly important role. In particular, improved models of sequence evolution accounting for heterogeneities across sites and across lineages have led to significant, although insufficient, improvement in phylogenetic accuracy. More recently, one main trend has been to move away from simple parametric models and stepwise approaches, towards integrative models explicitly considering the intricate interplay between multiple levels of macroevolutionary processes. Such integrative models are in their infancy, and their application to the phylogeny of eukaryotes still requires substantial improvement of the underlying models, as well as additional computational developments.

show abstract

“…The substitution model employed was LG+Γ, branch support was assessed by the rapid bootstrapping algorithm that is an inherent part of the best tree search strategy of RAxML. To test the robustness of the tree topologies we also employed ML inference using the program PhyML-CAT and the empirical profile mixture model C20 [35] with gamma correction (four categories) of the among-site rate heterogeneity; Chi2-based parametric branch support was calculated using the approximate likelihood ratio test implemented in PhyML (-b -2 option). Trees were visualized using iTOL (http://itol.embl.de/ [36]) and rendered for publication using a graphical editor.…”

Section: Alignment and Phylogenetic Analysesmentioning

confidence: 99%

Contrasting patterns in the evolution of the Rab GTPase family in Archaeplastida

Petrželková

Eliáš

2014

Acta Soc Bot Pol

View full text Add to dashboard Cite

Rab GTPases are a vast group of proteins serving a role of master regulators in membrane trafficking in eukaryotes. Previous studies delineated some 23 Rab and Rab-like paralogs ancestral for eukaryotes and mapped their current phylogenetic distribution, but the analyses relied on a limited sampling of the eukaryotic diversity. Taking advantage of the recent growth of genome and transcriptome resources for phylogenetically diverse plants and algae, we reanalyzed the evolution of the Rab family in eukaryotes with the primary plastid, collectively constituting the presumably monophyletic supergroup Archaeplastida. Our most important novel findings are as follows: (i) the ancestral set of Rabs in Archaeplastida included not only the paralogs Rab1, Rab2, Rab5, Rab6, Rab7, Rab8, Rab11, Rab18, Rab23, Rab24, Rab28, IFT27, and RTW (=Rabl2), as suggested previously, but also Rab14 and Rab34, because Rab14 exists in glaucophytes and Rab34 is present in glaucophytes and some green algae; (ii) except in embryophytes, Rab gene duplications have been rare in Archaeplastida. Most notable is the independent emergence of divergent, possibly functionally novel, in-paralogs of Rab1 and Rab11 in several archaeplastidial lineages; (iii) recurrent gene losses have been a significant factor shaping Rab gene complements in archaeplastidial species; for example, the Rab21 paralog was lost at least six times independently within Archaeplastida, once in the lineage leading to the "core" eudicots; (iv) while the glaucophyte Cyanophora paradoxa has retained the highest number of ancestral Rab paralogs among all archaeplastidial species studied so far, rhodophytes underwent an extreme reduction of the Rab gene set along their stem lineage, resulting in only six paralogs (Rab1, Rab2, Rab6, Rab7, Rab11, and Rab18) present in modern red algae. Especially notable is the absence of Rab5, a virtually universal paralog essential for the endocytic pathway, suggesting that endocytosis has been highly reduced or rewired in rhodophytes.

show abstract

Empirical profile mixture models for phylogenetic reconstruction

Cited by 305 publications

References 40 publications

Mutation-selection models of coding sequence evolution with site-heterogeneous amino acid fitness profiles

Mutation-selection models of coding sequence evolution with site-heterogeneous amino acid fitness profiles

Probabilistic models of eukaryotic evolution: time for integration

Contrasting patterns in the evolution of the Rab GTPase family in Archaeplastida

Contact Info

Product

Resources

About