cellHarmony: cell-level matching and holistic comparison of single-cell transcriptomes

摘要: To understand the molecular pathogenesis of human disease, precision analyses to define alterations within and between disease-associated cell populations are desperately needed. Single-cell genomics represents an ideal platform enable identification comparison normal diseased transcriptional populations. We created cellHarmony, integrated solution for unsupervised analysis, classification, types from diverse single-cell RNA-Seq datasets. cellHarmony efficiently accurately matches transcriptomes using a community-clustering alignment strategy compute differences in cell-type specific gene expression over potentially dozens Such used automatically identify distinct shared programs among cell-types impacted pathways regulatory networks impact perturbations at systems level. is implemented as python package workflow software AltAnalyze. demonstrate that has improved or equivalent performance alternative label projection methods, able likely cellular origins malignant states, stratify patients into clinical disease subtypes identified programs, resolve discrete impacting cell-types, illuminate therapeutic mechanisms. Thus, this approach holds tremendous promise revealing complex disease.