Research project Harnessing evolutionary transitions, machine learning, and genomics to decode pollen

Successful reproduction requires organisms to invest in a range of sexual traits. As a result, reproduction has a major impact on the way organisms look and function and on how their genomes evolve.

While negotiating reproduction is challenging, this task becomes even more of a challenge during evolutionary transitions, functional shifts with major effects on evolutionary processes. Evolutionary transitions change the rules of the game’ of reproduction, potentially sparking trait innovation and reshaping genomes. In particular, evolutionary transitions can alter the impact of sexual selection, selection which results from differential reproductive success due to variation in mating success. In animals, evolutionary transitions between internal and external fertilization have major effects on the rate and mode of sperm evolution, driven by dilution effects. In flowering plants, evolutionary transitions in pollination mode between insect and wind pollination should incur similar dilution effects, yet we lack robust tests of their impact on pollen evolution. Given widespread gene expression in pollen, shifts in sexual selection due to dilution effects should have marked morphological and genomic impacts, yet these effects remain incompletely characterized. By combining machine learning, computer vision, and state-of-the-art genomic approaches, this project will identify drivers and genomic consequences of pollen evolution in a broad comparative framework, across a range of evolutionary timescales.