Single-Cell Transcriptomics

Single-cell transcriptomics measures RNA in individual cells to reveal biological differences that bulk measurements can hide, helping researchers study development, disease, and treatment response. Common approaches include clustering and differential-expression analysis, integration of spatial and multi-omic data, gene regulatory network inference, and perturbation screens. Machine-learning methods range from latent-variable models and optimal-transport or flow-matching methods for trajectories and response distributions to Transformer foundation models for annotation, generation, and transfer. A central challenge is whether models generalize across donors, tissues, and unseen perturbations, and whether strong prediction scores capture causal biology rather than shared expression patterns; batch effects, limited data, and inconsistent evaluation complicate comparisons.