EvIL- Phylogenomics

Resolving the Tree of Life

Phylogenetic trees are an integral component of comparative analyses that span topics ranging from the transmission of SARS-CoV-2 to human cancer biology to the factors that have shaped patterns of uneven species richness across the planet. Understanding the phylogeny of major groups of organisms has attracted the attention of evolutionary biologists since Charles Darwin drafted his famous “I think” diagram in 1837. However, it took more than 100 years for Darwin’s abstract concept of the tree of life to develop into an explicit, data-driven approach to phylogeny reconstruction.

The early twenty-first century represents a phylogenetic renaissance. Over the past 20 years, hundreds of researchers have contributed resolved molecular phylogenies at varying time scales across the diversity of life. We continue to this effort to resolve all branches of the Tree of Life thereby provide a comparative framework for biologists. However, there are numerous pitfalls along the way, and we also develop computational means to avoid these

Avoiding Pitfalls

We employ quantitative measures from phylogenetic informativeness (PI) theory to provide a predictive framework with which to assess the influence of noise on phylogenetic datasets. The aims of this work are two-fold

First, by creating a predictive framework for phylogenomic studies, markers or types of sequence data can be chosen objectively, saving investigator time and money by enabling cost-effective research design. Second, by scrutinizing existing datasets, sources of conflicting information can be revealed, leading to more stable inferences

These approaches are incorporated into computational pipelines used in most of our research efforts, and are available in the R package PhyInformR. We continue making these methods approachable and site-rates for our anlyses can now be estimated in IQTree

On the birth and death of genes

The decay of phylogenetic information and ability to predict where to expect sequence convergence is directly linked to the ability to detect sequence homologs. We are currently expanding theory aimed at assessing loss of phylogenetic information to the problem of ortholog detection. This addresses a major challenge at deep evolutionary time scales or for rapidly evolving loci. Our goal is to provide an immensely powerful framework for disentangling functional orthologs from genetic orthologs and assessing the roles of adaptation and gene birth/death events in the maintenance of genetic diversity

Current Research Projects

Continue to resolve the ray-finned fish Tree of Life
The role of compositional bias and taxon sampling
Brachiobdellidan phylogenetics
Ortholog detection and gene family phylogeny
Experimental design for massive viral sequence datasets

Dataset Assembly and Missing Data

The enormous amount of freely available sequence data that continues to accumulate has tremendous promise for science. However, how to effectively harness this resource is not trivial. Large sequence datasets can be messy, and we often need to aggregate data from multiple sources. Moreover, high levels of missing data can impact our analyses. Extracting the data we need from within this tangled sea of information requires bioinformatic tools

To aid in this effort, we continue to develop TOAST, an R package that enables extraction of thousands of single copy orthologs for any taxonomic group using any combination of online or local data while also tracking missing data patterns. Building upon functions for assessing thresholds of acceptable missing data allows for custom data set creation, that work in concert with numerous utility functions that produce outputs directly for phylogenetic software such as BEAST or IQTREE. Our goal for TOAST is to provide readily accessible tools that can harvest any set of loci for any set of organisms with only a minimal knowledge of R required by end-users. This goal is in line with with our vision of empowering anyone to utilize public genomic scale datasets in their research

The Evolutionary Relationships of Ray-finned fishes

With over 30,000 species, ray-finned fishes comprise one out of every two vertebrates. While it is clear that distinct fishes such as pufferfishes, deep-sea anglerfishes, or seahorses form natural groups, the larger tree that unites these groups has eluded resolution until very recently

Dr. Dornburg worked with Dr. Thomas Near at Yale University to recently publish a review of the efforts to resolve ray-finned fish phylogeny. In many cases, the new ray- finned fish phylogeny radically departs from classical expectations based on morphology. Our review features scientific advances empowered by a broader understanding of ray-finned fish evolutionary history including implications for topics that span from the genetics of human health to reconsidering the concept of living fossils

We also discuss conceptual challenges that involve reconciling taxonomic classification with phylogenetic relationships and propose an alternate higher-level classification for Percomorpha, one of the most species-rich clades of fishes. Our review highlights remaining areas of phylogenetic uncertainty and opportunities for comparative investigations empowered by this new phylogenetic perspective on ray-finned fishes