Critiquing blind dating: the dangers of over-confident date estimates in comparative genomics

Abstract

Phylogenomic advances provide more rigorous estimates for the timing of evolutionary divergences than previously available (e.g., Bayesian relaxed-clock estimates with soft fossil constraints). However, because many family-level clades and higher, as well as model species within those clades, have not been included in phylogenomic studies, the literature presents temporal estimates likely harboring substantial errors. Blindly using such dates can substantially retard scientific advancement. We suggest a way forward by conducting analyses that minimize prior assumptions and use large datasets, and demonstrate how using such a phylogenomic approach can lead to significantly more parsimonious conclusions without a good fossil record. We suggest that such an approach calls for research into the biological causes of conflict between molecular and fossil signatures.