Optimal Rates for Phylogenetic Inference and Experimental Design in the Era of Genome-Scale Data Sets

Dornburg, Alex; Su, Zhuo; Townsend, Jeffrey P.

doi:10.1093/sysbio/syy047

Cited by 55 publications

(82 citation statements)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast to expectations of how markers from different genomic subsets are useful in phylogenetics studies (Dornburg, Su, and Townsend 2019;Steel and Leuenberger 2017;Phillips, Delsuc, and Penny 2004) , seven of the nine locus types showed broad phylogenetic utility over the timescales associated with this study. Intronic sites were moderately more informative for younger nodes in the rodent and combined datasets, while CDS sites had a stronger trend towards informativeness in older nodes among pecora, rodents, and the combined dataset.…”

Section: Phylogenetic Informativeness Of Most Locus Types Is Stable Acontrasting

confidence: 92%

“…homoplasy) over longer evolutionary timescales (Managadze et al 2011;Goldman and Yang 1994;Hughes and Yeager 1998;Bromham, Rambaut, and Harvey 1996;Ren et al 2016;Salichos and Rokas 2013;Dornburg et al 2014) . While this marker selection strategy appears reasonable, studies of the relationship between rate and phylogenetic utility are often complicated by interacting factors (Aguileta et al 2008;Townsend and Leuenberger 2011;Dornburg, Su, and Townsend 2019;Steel and Leuenberger 2017;Heath et al 2008;Su and Townsend 2015) .…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Genome-scale profiling reveals higher proportions of phylogenetic signal in non-coding data

Literman

Schwartz

2019

Preprint

View full text Add to dashboard Cite

Accurate estimates of species relationships are integral to our understanding of the evolution of traits and lineages, yet many relationships remain controversial despite whole genome sequence data. These controversies are due, at least in part, to complex patterns of phylogenetic and non-phylogenetic signal coming from regions of the genome experiencing different evolutionary forces. Here we profile the phylogenetic informativeness of loci from across mammalian genomes. We identified orthologous sequences from primates, rodents, and pecora, and annotated millions of sites as one or more of nine locus types (e.g. coding, intronic, intergenic) and profiled the informativeness of different locus types across the evolutionary timescales of each clade. In all cases, non-coding loci provided more overall signal and a higher proportion of phylogenetic signal compared to coding loci. Most locus types provide relatively consistent phylogenetic information across timescales, although we find evidence that coding and intronic regions may inform disproportionately about older and younger splits, respectively, to a limited degree. We also validated the SISRS pipeline as an annotation-free ortholog discovery pipeline that identifies millions of phylogenetically informative sites directly from whole-genome sequencing reads.

show abstract

Section: Phylogenetic Informativeness Of Most Locus Types Is Stable Acontrasting

confidence: 92%

Section: Introductionmentioning

confidence: 99%

Genome-scale profiling reveals higher proportions of phylogenetic signal in non-coding data

Literman

Schwartz

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…While no studies have directly examined the impact of model complexity on time estimation in phylogenomic investigations, there have been reports that simple substitution models often perform similarly to or slightly worse than complex substitution models in some types of phylogenetic inferences (Tamura et al 2004;Yoshida and Nei 2016;Dornburg et al 2018;Spielman and Kosakovsky Pond 2018;Abadi et al 2019;Spielman 2019). However, in molecular dating, it is intuitively assumed that underestimation of sequence divergence caused by the use of simple models will result in significantly distorted time estimates.…”

Section: Introductionmentioning

confidence: 99%

Relative efficiencies of simple and complex substitution models in estimating divergence times in phylogenomics

Tao

Barba-Montoya

Huuki

et al. 2020

Preprint

View full text Add to dashboard Cite

The conventional wisdom in molecular evolution is to apply parameter-rich models of nucleotide and amino acid substitutions for estimating divergence times. However, the actual extent of the difference between time estimates produced by highly complex models compared to those from simple models is yet to be quantified for contemporary datasets that frequently contain sequences from many species and genes. In a reanalysis of many large multispecies alignments from diverse groups of taxa using the same tree topologies and calibrations, we found that the use of the simplest models can produce divergence time estimates and credibility intervals similar to those obtained from the complex models applied in the original studies. This result is surprising because the use of simple models underestimates sequence divergence for all the datasets analyzed. We find three fundamental reasons for the observed robustness of time estimates to model complexity in many practical datasets. First, the estimates of branch lengths and nodeto-tip distances under the simplest model show an approximately linear relationship with those produced by using the most complex models applied, especially for datasets with many sequences. Second, relaxed clock methods automatically adjust rates on branches that experience considerable underestimation of sequence divergences, resulting in time estimates that are similar to those from complex models. And, third, the inclusion of even a few good calibrations in an analysis can reduce the difference in time estimates from simple and complex models. The robustness of time estimates to models complexity in these empirical data analyses is encouraging, because all phylogenomics studies use statistical models that are oversimplified descriptions of actual evolutionary substitution processes.

show abstract

“…For example, we and others have demonstrated that employing the most parameter-rich model, GTR+I+G, leads to topology reconstructions that are as accurate as those produced by established model selection strategies (Arbiza et al 2011;Abadi et al 2019;Spielman 2019). On the other hand, for the tasks of branch-length and divergence time estimations it was shown that the accuracy is conditioned on the selected model and so model selection should still be advocated (Buckley et al 2001;Posada 2001;Minin et al 2003;Abdo et al 2005;Spielman and Kosakovsky Pond 2018;Dornburg et al 2019). For example, Abadi et al (Abadi et al 2019) demonstrated that the consistent use of GTR+I+G generally results in inferior branch-length estimates compared to those obtained by model selection criteria, with the BIC and DT criteria obtaining the most accurate estimates.…”

Section: Introductionmentioning

confidence: 98%

ModelTeller: model selection for optimal phylogenetic reconstruction using machine learning

Abadi

Avram

Pupko

et al. 2020

Preprint

View full text Add to dashboard Cite

Statistical criteria have long been the standard for selecting the best model for phylogenetic reconstruction and downstream statistical inference. While model selection is regarded as a fundamental step in phylogenetics, existing methods for this task consume computational resources for long processing time, they are not always feasible, and sometimes depend on preliminary assumptions which do not hold for sequence data. Moreover, while these methods are dedicated to revealing the processes that underlie the sequence data, in most cases they do not produce the most accurate trees. Notably, phylogeny reconstruction consists of two related tasks, topology reconstruction and branch-length estimation. It was previously shown that in many cases the most complex model, GTR+I+G, leads to topologies that are as accurate as using existing model selection criteria, but overestimates branch lengths. Here, we present ModelTeller, a computational methodology for phylogenetic model selection, devised within the machine-learning framework, optimized to predict the most accurate model for branch-length estimation accuracy. ModelTeller relies on a readily implemented machine-learning model and thus the prediction according to features extracted from the sequence data results in a substantial decrease in running time compared to existing strategies. We show that on datasets simulated under simple homogenous substitution models ModelTeller leads to branch-length estimation that is as accurate as the statistical model selection criteria. We then demonstrate that ModelTeller outperforms these criteria when more intricate patterns -that aim at mimicking realistic processes -are considered.

show abstract

Optimal Rates for Phylogenetic Inference and Experimental Design in the Era of Genome-Scale Data Sets

Cited by 55 publications

References 74 publications

Genome-scale profiling reveals higher proportions of phylogenetic signal in non-coding data

Genome-scale profiling reveals higher proportions of phylogenetic signal in non-coding data

Relative efficiencies of simple and complex substitution models in estimating divergence times in phylogenomics

ModelTeller: model selection for optimal phylogenetic reconstruction using machine learning

Contact Info

Product

Resources

About