Poster on Single cell Causal Analysis of cell fate transition via in silico genetic PErturbation (SCAPE).
Date:
Dissecting the causal effects of complex gene regulatory interactions is essential for predicting cell fate, particularly during transitions induced by genetic perturbations. In this study, we introduce SCAPE, a deep learning-based method for modeling transcription dynamics that predicts the post-perturbation transcriptomic velocity field. This allows SCAPE to effectively predict cell fate changes resulted from genetic alterations in specific genes. We validated SCAPE’s performance using both simulated and real datasets, including those related to neurogenesis, gastrulation, pancreatic endocrinogenesis, endothelial-to-hematopoietic transition, and spermiogenesis. SCAPE demonstrates superior performance compared to competing methods in accurately recapitulating well-characterized cell state transitions induced by genetic perturbations. We anticipate that SCAPE will enhance our understanding of cell fate transitions and facilitate the discovery of key regulators in developmental processes and diseases.