Bayesian phylodynamics of early vertebrate development in BEAST 2

Analysing single-cell lineage relationships of an organism is crucial towards understanding the fundamental cellular dynamics that drive development. CRISPR-based dynamic lineage tracing relies on recent advances in genome editing and sequencing technologies to generate inheritable, evolving genetic barcode sequences which enable reconstruction of such cell lineage trees, also referred to as phylogenetic trees. Recent work generated custom computational strategies to produce robust tree estimates from such data. We further capitalise on these advancements and introduce GABI (GESTALT Analysis using Bayesian Inference), which extends the analysis of GESTALT (Genome Editing of Synthetic Target Arrays for Lineage Tracing) data to a fully integrated Bayesian phylogenetic inference frame-work in the software BEAST 2. This implementation allows to represent the uncertainty in reconstructed tree reconstruction and enables their scaling in absolute time. Furthermore, based on such time-scaled lineage trees, the underlying processes of growth, differentiation and apoptosis are quantified through so-called phylodynamic inference, typically relying on a birthdeath or coalescent model. After validating the implementation, we demonstrate that the methodology results in robust estimates of lineage trees and growth dynamics characteristics of early zebrafish Danio rerio development. GABIs codebase is publicly available at https://github.com/azwaans/GABI.