Gillian Chu will present her General Exam "Probabilistic Models for Evolutionary Tree Reconstruction" on Monday, April 22, 2024 at 3:00 PM in Friend 202. Committee Members: Ben Raphael (advisor), Mona Singh, Yuri Pritykin Abstract: Lineage tracing is the tracking of cell divisions from a single cell to a resulting multicellular organism. Recent experimental work in developmental biology has resulted in a new generation of dynamic lineage tracing technologies which edit a cell to accumulate heritable CRISPR edits over generations. The subsequent computational problem of reconstructing the cell lineage tree from such dynamic lineage tracing data has also been the subject of much recent work. However, several unaddressed computational challenges remain. For instance, one key unaddressed challenge is how to handle the high rates (up to 40%) of heterogeneous (heritable silencing and stochastic dropout) missing data. In this work, we address these challenges by presenting a novel probabilistic approach: (a) we propose a new model of dynamic lineage traced sequence evolution with probabilistic mixed-type missing (PMM) data, and (b) present Lineage Analysis via Maximum Likelihood (LAML), which infers the (i) cell phylogeny with (ii) time-scaled branch lengths and (iii) missing data rates under this model. Our work demonstrates a systematic bias in previous mutation-based approaches to branch length estimation, with important implications for downstream analyses. Crucially, using the LAML time-scaled tree we can place evolutionary events (such as cancer cell migration to and from different anatomical locations) in experimental time. From a lineage-traced mouse model of metastasis, we observe three “epochs” of metastasis with distinct cell migration patterns. Reading List: https://docs.google.com/document/d/1Pjmf-OA0QhbapMoPxzRP62_ukFQ6GfCaKeIOjBFC... Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so.