ClockstaRX: testing molecular clock hypotheses with genomic data

Abstract

Phylogenetic studies of genomic data can provide valuable opportunities for evaluating evolutionary timescales and drivers of rate variation. These analyses require statistical tools based on molecular clocks. We present ClockstaRX, a flexible platform for exploring and testing evolutionary rate signals in phylogenomic data. It implements methods that use information from gene trees in Euclidean space, allowing data transformation, visualization, and hypothesis testing. ClockstaRX implements formal tests of the dimensionality reducibility of the Euclidean space of rates, and for identifying loci and branches that have a large influence on rate variation. Using simulations to evaluate the performance of the methods implemented, we find that inferences about rates can be strongly influenced by the overall amount of rate variation in the data, the shared patterns of among-lineage rate heterogeneity across groups of loci, and missing data. In an analysis of phylogenomic data from birds, we find a higher rate of evolution in introns compared with exons across all lineages. In addition, passerine taxa are highlighted as having unique patterns of genomic evolutionary rates compared with other avian lineages. Drawing on these results, we recommend careful exploratory analyses and filtering before performing phylogenomic analyses using molecular clocks.