Species Tree Inference on Data with Paralogs is Accurate Using Methods Intended to Deal with Incomplete Lineage Sorting

Abstract

Received ...; reviews returned ...; accepted ...Abstract.---The multispecies coalescent (MSC) has emerged as a powerful and desirable framework for species tree inference in phylogenomic studies. Under this framework, the data for each locus is assumed to consist of orthologous, single-copy genes, and heterogeneity across loci is assumed to be due to incomplete lineage sorting (ILS). These assumptions have led biologists that use ILS-aware inference methods, whether based directly on the MSC or proven to be statistically consistent under it (collectively referred to here as MSC-based methods), to exclude all loci that are present in more than a single copy in any of the studied genomes. Furthermore, such analyses entail orthology assignment to avoid the potential of hidden paralogy in the data. The question we seek to answer in this study is: What happens if one runs such species tree inference methods on data where paralogy is present, in addition to or without ILS being present? Through simulation studies and analyses of two biological data sets, we show that running such methods on data with paralogs provide very accurate results, either by treating all gene copies within a family as alleles from multiple individuals or by randomly selecting one copy per species. Our results have significant implications for the use of MSC-based phylogenomic analyses, demonstrating that they do not have to be restricted to single-copy loci, thus greatly increasing the amount of data that can be used. [Multispecies coalescent; incomplete lineage sorting; gene duplication and loss; orthology; paralogy.]

Polynomial-Time Statistical Estimation of Species Trees under Gene Duplication and Loss

“…Their study has an advantage over our study in that it explored larger datasets (up to 500 species), but they did not consider ASTRAL, and all their genes evolved under a strict molecular clock. Another study [10] evaluated ASTRAL-multi on true gene trees, and found it had very high accuracy. Overall, our study is the first study to explore ASTRAL-multi on estimated gene trees.…”

“…We evaluated methods on simulated datasets based on the 16-taxon fungal dataset studied in [10,25]. We used this fungal dataset to construct a model species tree, on which we generated gene trees using SimPhy [20] under three GDL rates (the lowest of which is the one selected by [25] to reflect the fungal dataset, so that the higher two reflect more challenging conditions).…”

“…ASTRAL We provide theoretical guarantees and empirically validate an approach based on ASTRAL [21] in its variant for multiple alleles [24], which we refer to as ASTRAL-multi. Following [10], the input consists of unrooted MULtrees T from all gene families, where copies of a gene in a species are treated as multiple alleles within the species.…”

“…As a direct consequence of our identifiability proof, it is straightforward to establish the statistical consistency of the following pipeline, which we refer to as ASTRAL/ONE (see also [10]): for each gene tree t i , pick in each species a random gene copy (if possible) and run ASTRAL on the resulting set of modified gene treest i . The proof can be found in the appendix.…”

“…Here we describe the simulation of gene trees from a species tree under a model of GDL. Our protocol is based on the simulation study by [10] that uses a biological datasets (16 species of fungi) from [25]. We simulated gene trees with different rates of GDL from the species tree estimated by Rasmussen and Kellis [25] (download: http://compbio.mit.edu/dlcoal/pub/config/fungi.stree); note the estimated species tree has branch lengths in millions years (myr) and the estimated number of generations is 10 per year.…”