Abstract
Analyzing the trajectories of cellular differentiation sheds light on key questions across biology, from how cell types are stabilized during embryonic development to how they destabilize with age or disease. New single cell measurement technologies offer the prospect of reconstructing these developmental trajectories from snap-shots of cell state together with lineage trees inferred from continuously-induced mutations in heritable DNA-barcodes. However, current methods for reconstructing developmental trajectories are not designed to leverage the additional information contained in lineage trees. Inspired by these recent experimental advances, we present a novel framework for reconstructing developmental trajectories from snapshots of cell state combined with lineage trees. Our method learns from both kinds of information together using mathematical tools from graphical models and optimal transport. We find that lineage data helps disentangle complex state transitions with fewer measured time points, enabling increased accuracy for lower experimental cost. Moreover, integrating lineage tracing with trajectory inference in this way enables accurate reconstruction of developmental pathways that are impossible to recover with state-based methods alone.
Competing Interest Statement
The authors have declared no competing interest.