Prioritizing prognostic-associated subpopulations and individualized recurrence risk signatures from single-cell transcriptomes of colorectal cancer

Abstract

Colorectal cancer (CRC) is one of the most common gastrointestinal malignancies. There are few recurrence risk signatures for CRC patients. Single-cell RNA-sequencing (scRNA-seq) provides a high resolution platform for prognostic signature detection. However, scRNA-seq is not practical in large cohorts due to its high cost and most single-cell experiments lack clinical phenotype information. Few studies have been reported to use external bulk transcriptome with survival time to guide the detection of key cell subtypes in scRNA-seq data. We proposed a data analysis framework to prioritize prognostic-associated subpopulations based on relative expression orderings (REOs). Cell type specific gene pairs (C-GPs) were identified to evaluate prognostic value for each cell type. We found REOs-based signatures could accurately classify most cell subtypes. C-GPs achieves higher precision compared with four current methods. Moreover, we developed single-cell gene pair signatures to predict recurrence risk for patients individually. Fibro_SGK1 cells and IgA+ IGLC2+ B cells were novel prognostic-associated subpopulations. A user-friendly toolkit, scRankXMBD (https://github.com/xmuyulab/scRank-XMBD), was developed to enable implementation of this framework. Our work facilitate the application of the rank-based method in scRNA-seq data for prognostic biomarker discovery and precision oncology.