Integrated analysis of fossils and molecular divergence time estimation infers a latest Jurassic origin of angiosperms

Molecular timescales are based on the calibration of molecular evolution to geological time using fossil constraints, but conventional calibration strategies use limited and often subjectively interpreted fossil data. Here, we used the Bayesian Brownian Bridge model to derive data-driven calibration densities based on extensive fossil occurrence records. This approach can integrate the uncertainty on extant and historical diversity to estimate clade age. We transformed the estimated ages based on >25,000 fossil occurrences into calibration densities, which we used to constrain 110 node ages in a 644-species angiosperm phylogenetic tree inferred from a molecular alignment of 83 genes. The results refute a post-Jurassic origin of angiosperms, suggesting a cryptic early history. Our study demonstrates the utility of a mechanistic approach to establish node age constraints in molecular clock-dating analyses, resulting in a more objective method to integrate molecular and palaeontological data when inferring evolutionary timescales.