Machine Learning for Evolutionary Genomics Data

Evolutionary genomics and population genetics investigate patterns of genetic diversity between species or between populations within a species and play a fundamental role in many aspects, from theoretical facets of evolution to practical ones, such as conservation genetics and biomedical sciences. Methodologies have always been a strong interest of the community, from the development of mathematical models to the design of statistical inference tools, leading to numerous biological discoveries. These developments helped to adapt very quickly to the continuous influx of data, which has not only dramatically increased in quantity but also keeps changing in terms of quality and type. Among the methodological frameworks, machine learning has emerged as a promising way of analysing large and complex datasets. The application of AI, and particularly deep learning, to evolutionary genomics is still in its infancy while showing promising initial results. It is currently applied to a variety of tasks, such as the inference of demographic history, ancestry, natural selection, phylogeny, species delimitation and diversification. However, machine learning methods in evolutionary genomics and population genetics face unique challenges, including identifying appropriate assumptions about the evolutionary process and how to simulate it, and identifying the best ways to handle sequences, sequence alignments, phylogenetic trees, or additional information, such as geographical maps, temporal labels and environmental covariates. Overcoming these challenges requires a collaborative effort. The goal of this conference is to foster this effort by allowing interested researchers to meet and share their work.