Natsidis, Paschalis; (2022) Systematic errors in phylogenomics with a focus on the major metazoan clade Deuterostomia. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Natsidis_10148450_Thesis_sig-removed.pdf Download (17MB) | Preview

Abstract

Modern-day phylogenomics studies employ large data sets of many genes to resolve evolutionary relationships among many species. A typical phylogenomic workflow consists of certain steps: taxon sampling, orthology inference, marker selection and tree search. All of these steps contain some subjective decisions made by the researcher, posing risks for introducing systematic errors in the final results. In this thesis, I investigate the source and the impact of systematic errors in multiple steps of the phylogenomic workflow, focusing on the major clade Metazoa. First, I create simulated sets of orthologs under different settings for evolutionary rate and rate heterogeneity among sites and use OrthoFinder to infer their (known) orthology relationships. Orthology inference is sensitive to high evolutionary rates and low rate heterogeneity among sites. I show that errors in orthology inference are carried over to downstream analysis such as gene presence/absence phylogenies, gene gains/losses inference and phylostratigraphy. I also introduce a novel computational pipeline which allows us to identify the presence of a hidden break in the 28S ribosomal RNA of a given species. Mapping RNA-seq reads onto the 28S rRNA sequence reveals non-existent coverage of mapped reads near the middle of the 28S rRNA sequence of species that possess the hidden break. I apply this pipeline in hundreds of metazoan and other eukaryotic species and find that the hidden break is a rarely lost protostome feature, with surprising events of convergent evolution outside Metazoa. I finally focus on the major metazoan clade of Deuterostomia; while it has been widely accepted as a monophyletic group for over a century, recent phylogenomic studies addressing known systematic errors have recovered low support for monophyletic Deuterostomia. I examine five recently published metazoan phylogenomic data sets to show that monophyletic Deuterostomia is much less well supported than monophyletic Protostomia. I also create 40 new data sets, with and without fast-evolving taxa, and use them to correlate strong support for monophyletic Deuterostomia with problematic conditions in a phylogenomic analysis.

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item