Learning evolutionary parameters from B-cell lineage trees.

Characterising selection and its impact on genetic diversity within evolving populations is a fundamental challenge in evolutionary biology. B-cell affinity maturation, a process exemplifying accelerated evolution under selection, offers a unique opportunity to investigate these dynamics within shorter timescales. Despite its significance, current methods face limitations: they often do not extend beyond neutrality, and there is a lack of reliable summary statistics for discerning non-neutral evolutionary processes. To address this issue, we construct mathematical models that capture the dynamics of proliferating cell populations undergoing non-neutral mutations, and, informed by these models we develop an expectation maximisation inference scheme to learn the parameters of selection from data. We will show progress towards applying the algorithm to B-cell repertoire data.