Whole genome phylogeny of Gallus: introgression and data-type effects

Abstract

Background: Previous phylogenetic studies that include the four recognized species of Gallus have resulted in a number of distinct topologies, with little agreement. Several factors could lead to the failure to converge on a consistent topology, including introgression, incomplete lineage sorting, different data types, or insufficient data. Methods:We generated three novel whole genome assemblies for Gallus species, which we combined with data from the published genomes of Gallus gallus and Bambusicola thoracicus (a member of the sister genus to Gallus). To determine why previous studies have failed to converge on a single topology, we extracted large numbers of orthologous exons, introns, ultra-conserved elements, and conserved non-exonic elements from the genome assemblies. This provided more than 32 million base pairs of data that we used for concatenated maximum likelihood and multispecies coalescent analyses of Gallus.Results: All of our analyses, regardless of data type, yielded a single, well-supported topology. We found some evidence for ancient introgression involving specific Gallus lineages as well as modest data type effects that had an impact on support and branch length estimates in specific analyses. However, the estimated gene tree spectra for all data types had a relatively good fit to their expectation given the multispecies coalescent. Conclusions:Overall, our data suggest that conflicts among previous studies probably reflect the use of smaller datasets (both in terms of number of sites and of loci) in those analyses. Our results demonstrate the importance of sampling large numbers of loci, each of which has a sufficient number of sites to provide robust estimates of gene trees. Low-coverage whole genome sequencing, as we did here, represents a cost-effective means to generate the very large data sets that include multiple data types that enabled us to obtain a robust estimate of Gallus phylogeny.

Data Types and the Phylogeny of Neoaves

“…The tight linkage between coding and non-coding in the Prum dataset makes this study an excellent complement to Reddy et al [13], which compared trees estimated using unlinked coding and non-coding data. When the evidence for data type effects in this study is combined with the results of Reddy et al [13] and Jarvis et al [11] they provide strong evidence that the important variable is the data types and not any idiosyncratic features of specific genomic regions in each study, Data type effects have been described in a number of studies, though the nature of the effects range from those that are quite subtle [49,53] to much stronger effects [11,13,[18][19][20][21][22][23]. Some reported examples of data type effects reflect analyses of the same coding regions as nucleotides and after translation to amino acids [18,[54][55][56].…”

“…For this. study define data type effects in a simple and empirical manner: data type effects are those cases where phylogenetic analyses of distinct subsets of the genome defined using non-phylogenetic criteria (e.g., structural or functional criteria) yield different results [1,13,24,49]. It is necessary to restrict consideration to data types dispersed across the genome so discordant histories for individual genes [50,51] do not obscure the data type dependence (if it exists).…”

Deep-Time Demographic Inference Suggests Ecological Release as Driver of Neoavian Adaptive Radiation

“…Herein, we combine estimates of CBL obtained using genome-scale indel data with information from timetrees and reconstructions of ancestral generation times to yield what we believe to be the first empirical estimates of the relative effective population size (Ne of ancestral lineages in very deep time. Our approach bears some similarity to recently described approaches [81,82] and those of Tiley et al [83], who estimated CBL using binary data (TE insertions and "perfect" transversions, respectively). However, we implemented these analyses in a modified version of the commonly used multispecies coalescent program Accurate Species TRee Algorithm (ASTRAL) [84,85].…”

“…Alternatively, one might estimate the timetree using loci selected from parts of the genome where Ne is reduced locally. Tiley et al [83] reported that conserved non-exonic elements (from Edwards et al [114]) had shorter internal branch lengths than any other data type they examined. They proposed that this reflects a reduction in SPILS due to local reductions in Ne driven by Hill-Robertson interference [115].…”

“…The popular expectation that Ne is not stable may be rooted in the fact that most studies of Ne in natural populations are biased in focus toward species of conservation concern that have experienced population bottlenecks. Signatures of ILS in recently diverged species are well documented (e.g., [83,121]) and are calculated to persist as long as 10 million years [122]. One important consideration for indels as a marker of ILS is that they are not subject to recombination.…”