Intrinsic dimensionality of single-cell transcriptomic data reveals potency landscapes during cell reprogramming

Abstract

Cell potency, the ability of a cell to generate other cell types, is a fundamental property that drives development, regeneration, and reprogramming. Recent computational advances have enabled estimation of potency directly from single-cell RNA sequencing (scRNA-seq) data, with intrinsic dimensionality (ID) emerging as a promising, data-driven measure of transcriptional complexity linked to developmental potential. While ID offers an unbiased and interpretable framework, existing implementations have two main limitations: they have been applied in only a narrow range of biological contexts, and rely on clustering, which restricts resolution at the single-cell level. Here, we introduce IDEAS (Intrinsic Dimensionality Estimation Analysis of single-cell RNA sequencing data), a Python-based toolkit that computes both global and single-cell ID from scRNA-seq data, enabling potency scoring without requiring predefined clusters. IDEAS extends ID-based potency estimation to single cells, showing its usefulness in a new biological setting. It offers a clear and reliable way to study cell plasticity, simplifies ID score calculation, and makes the method easier to use across different biological systems, speeding up research on this approach to measuring cell potency. We apply IDEAS to multiple datasets of cellular reprogramming, a dynamic and heterogeneous process in which cells transition between identities. Our analysis reveals that ID scores effectively capture changes in potency, identify partially reprogrammed intermediates, and delineate alternative reprogramming trajectories.