scTOP: physics-inspired order parameters for cell fate classification and visualization of single cells

Recent advances in single-cell RNA-sequencing (scRNA-seq) and lineage tracing techniques provide an unprecedented window into the biology of cellular identity. The wealth of data calls for new theoretical and computational frameworks for understanding cell fate specification, accurately classifying cell fates from expression data, and integrating knowledge from cell atlases. Here, we present single-cell Type Order Parameters (scTOP): a statistical-physics-inspired approach for constructing "order parameters" for cell fate given a reference basis of cell types. scTOP achieves near state-of-the-art performance for cell identification at the resolution of single cells and yields interpretable visualizations of developmental trajectories, such as bifurcations between closely related cell fates. Importantly, scTOP does this without using feature selection, statistical fitting, or dimensional reduction (e.g. UMAP, t-SNE, PCA, SPRING). We illustrate the power of scTOP on a wide variety of human and mouse datasets (both in vivo and in vitro), including existing tissue atlases and lineage tracing data. We also provide an easy-to-use Python package implementation of scTOP. Our results suggest that physics-inspired order parameters can serve as an important tool for understanding and analyzing developmental landscapes and cellular identity across biological contexts and organisms.