Phylogenonomics of cockroaches & termites
My research has contributed some of the first nuclear genomic studies of cockroach phylogenetics. In Proceedings of the Royal Society B, we integrated a transcriptome-based phylogeny of cockroaches with divergence dating analysis and numerous other analyses to come to novel insights about the evolution of cockroaches and termites. This data-set demonstrated strong statistical support for previously hypothesized relationships. In particular, we demonstrated that lamproblattid cockroaches are the closest relative to the social and wood-feeding Cryptocercus cockroaches and termites. Establishing this relationship provided a basis for coming to the conclusion that wood-burrowing/feeding evolved, in part, to relieve predatory pressure on egg cases (Part B in figure to left). We also integrated this phylogenetic framework with other phenotypic data to reconstruct various aspects of cockroach evolution (panels A-F on figure to left).
In this research we also spent great efforts in revising the fossil record to assess the timing of important events in the evolution of Blattodea. As a result of this massive genetic data set and a largely revised fossil record, we inferred that most events in cockroach and termite evolution occurred much earlier than previously thought−significantly bridging gaps in the fossil record. This dating analysis corroborates suspicions that abundant "cockroach-like" fossils in the Carboniferous could not have been crown-Blattodea. Furthermore, it limits the occurrence of termites to the Cretaceous period−discrediting older purported fossil evidence of termites.
Phylogenetic experimental design
In my NSF funded postdoc, and my postdoc with at UTK I am using custom designed next-gen DNA data (see below) sequenced from Blaberoidean cockroaches to explore ways in which phylogenetic inference can be generally improved.
First, although many researchers have access to millions of nucleotides of sequence data, utilizing all this data for phylogenetic inference can be problematic. Sequence data can contain noise that is unevenly distributed. Signal for certain relationships can be rare and may only be found in a handful of loci. Thus, reducing datasets to their most informative elements could improve phylogenetic accuracy and precision, while also yielding shorter computation times and other practical improvements.
Second, coalescent approaches to species-tree inference (which have become increasingly popular) rely strongly on gene-tree topologies. Yet, limited sequence information in single genes may be inadequate for accurate gene-tree inference. To account for this, I am applying novel models of evolution that utilize information content nested in multiple scales (i.e. nucleotide, codon, protein) to test general trends among estimated gene trees and the implications for coalescent species tree inference.
In all these projects, I utilize super-family Blaberoidea as a study system. These cockroaches are extremely species rich among Polyneopteran insects (second only to grasshoppers and crickets). To date, the relationships among the major lineages of this group have only been investigated superficially. Previous investigations have shown that this clade has many "problematic" regions, including at least one rapid radiation and a variety of "long branch" taxa. Improving upon our understanding of their phylogeny can help resolve the pattern of evolution of many interesting traits that are currently unknown. In particular, multiple independent switches in genital symmetry may have been important events preceding rapid radiations, and these may correspond to invasions of new biogeographical zones as well.
This is an expansion of the 1KITE project.
I also recently published a phylogeny of cockroaches using few loci and hundreds of species. In this study we demonstrated that support for relationships in this phylogeny can only be meaningfully demonstrated using multiple approaches in parallel.
In conjunction with the landscape ecology project we wanted to find out the extent to which certain landscape elements were affecting population level evolution in the Rupununi savanna. We hypothesized that savannas and flood regions were isolating populations. Students in a Biotechnology class at Peddie School in Hightstown, NJ assisted with this project by collecting preliminary mitochondrial data. They showed that gene flow was abundant among populations separated by up to 8 miles of savanna or floodland. Consequently, these students provided some of the only evidence ever demonstrating cockroach dispersal abilities.
Given the highly polymorphic nature of cockroaches (age stage and sexual polymorphism) and the difficulty in using taxonomic literature, genetic barcoding provides a simple alternative to traditional identification techniques.
We have used genetic barcodes for age stage and sex association in taxonomic work, morphotype definitions in our richness study, and identifications of taxa.
Functional gene evolution and transcriptomics
In collaboration with the 1KITE (1K Insect Transcriptomes) sub-group for Dictyoptera I am analyzing 100 transcriptomes of cockroaches, mantises and termites to examine functional gene evolution that may have led to rapid diversifications and the evolution of eusociality in termites.