Unveiling Clonal Cell Fate and Differentiation Dynamics: A Hybrid NeuralODE-Gillespie Approach

Abstract

Recent lineage tracing single-cell techniques (LT-scSeq), e.g., the Lineage And RNA RecoverY (LARRY) barcoding system, have enabled clonally resolved interpretation of differentiation trajectories. However, the heterogeneity of clone-specific kinetics remains understudied, both quantitatively and in terms of interpretability, thus limiting the power of bar-coding systems to unravel how heterogeneous stem cell clones drive overall cell population dynamics. Here, we present CLADES, a NeuralODE-based framework to faithfully estimate clone-specific kinetics of cell states from newly generated and publicly available human cord blood LARRY LT-scSeq data. By incorporating a stochastic simulation algorithm (SSA) and differential expression gene (DEGs) analysis, CLADES yields cell division dynamics across differentiation timecourses and fate bias predictions for the early progenitor cells. Moreover, clone-level quantitative behaviours can be grouped into characteristic types by pooling individual clones into meta-clones. By benchmarking with CoSpar, we found that CLADES improves fate bias prediction accuracy at the meta-clone level. In conclusion, we report a broadly applicable approach to robustly quantify differentiation kinetics using meta-clones while providing valuable insights into the fate bias of cellular populations for any organ system maintained by a pool of heterogeneous stem and progenitor cells.