Assessment of methods for amino acid matrix selection and their use on empirical data shows that ad hoc assumptions for choice of matrix are not justified

Abstract

Background: In recent years, model based approaches such as maximum likelihood have become the methods of choice for constructing phylogenies. A number of authors have shown the importance of using adequate substitution models in order to produce accurate phylogenies. In the past, many empirical models of amino acid substitution have been derived using a variety of different methods and protein datasets. These matrices are normally used as surrogates, rather than deriving the maximum likelihood model from the dataset being examined. With few exceptions, selection between alternative matrices has been carried out in an ad hoc manner.

Yeast forms dominate fungal diversity in the deep oceans

“…These analyses demonstrated no candidate chimaera sequences. The final masked alignment was analysed using MODELGEN-ERATOR (Keane et al 2004) to identify the most appropriate model for phylogenetic analyses (GTRCG (four discrete categories, aZ0.57)Cproportion of invariant sites (0.49)). Then, the phylogeny was calculated using two methods: (i) PHYML (Guindon et al 2005) using the model parameters estimated using MODELGENERATOR and 100 bootstrap replicates, and (ii) MRBAYES3 (Ronquist & Huelsenbeck 2003) analysis was conducted for 1 000 000 generation samples using model of site rate variation as mentioned above but allowing the MCMCMC to search alternative model parameter values.…”

The ring of life hypothesis for eukaryote origins is supported by multiple kinds of data

“…However, if evolutionary history of the evolving objects under study has contained introgressive events, then a simple diverging tree cannot be used. To use a tree when the data were not generated by a tree-like process is an error in model selection [92]. There are many network-like alternatives to trees, including phylogenetic networks [93], sequence-sharing networks [32] and N-rooted fusion networks [31,33].…”

Mitochondrial genomes reveal the extinct Hippidion as an outgroup to all living equids

“…For PHYML analyses, we merged these partitions into one alignment, as PHYML does not handle multiple partitions. The best substitution model for each partition/merged dataset (with and without outgroups) was determined with MODELGENERATOR v. 0.85 [20] on the basis of its Bayesian information criterion, and applied to the corresponding partition/dataset in BEAST (unlinked substitution models) and PHYML analyses (electronic supplementary material, tables S6-S7). BEAST analyses were performed assuming a birth-death serially sampled tree model, and divergence dates were estimated using a log-uncorrelated molecular clock model together with tip-sampling (electronic supplementary material, table S8) and assuming a time for the MRCA at 4.0-4.5 Ma for Equus [2].…”