Inferring the evolutionary history of the Sino-Himalayan biodiversity hotspot using a Bayesian birth–death skyline model

Abstract

The current status of the Sino-Himalayan region as a biodiversity hotspot, particularly for flora, has often been linked to the uplift of the Sino-Tibetan Plateau and Himalayan and Hengduan Mountains. However, the relationship between the topological development of the region and the onset of diversification is yet to be confirmed. Here, we apply Bayesian phylodynamic methods to a large phylogeny of angiosperm species from the Sino-Himalayas to infer changes in their rates of evolution through time. We find strong evidence for high diversification rates in the Paleocene, late Miocene and Pliocene, and for negative diversification rates in the Quaternary driven by an increase in extinction rates. Our analyses suggest that changes in global palaeotemperatures are unlikely to be a driving force for these rate shifts. Instead, the collision of the Indian continent with Eurasia and coeval topographic change in the Sino-Himalayas, the Miocene Grassland Expansion and the impact of Pleistocene glaciations on this altitudinally variable region may have driven these rates. We also demonstrate the strong influence of the choice of change times on the shape of inferred piecewise-constant trajectories in Bayesian phylodynamics, and advocate for the use of prior information when making this decision.