Integration of single-cell and bulk RNA-sequencing data to construct and validate a signature based on NK cell marker genes to predict immunotherapy response and prognosis in colorectal cancer

We aimed to create a NK cell marker genes-based signature to predict immunotherapy response and prognosis in colorectal cancer. We integrated scRNA-seq data from four Gene Expression Omnibus (GEO) samples and performed Weighted gene correlation network analysis (WGCNA) based on 587 the Cancer Genome Atlas (TCGA) colorectal cancer samples to uncover NK cell-related genes. We identified 1080 NK cell-related core genes and 276 NK cell-related feature genes based on WGCNA and clustering and annotation of scRNA-seq data, respectively. Six key NK cell-related prognostic signature genes were obtained by univariate and LASSO regression analyses, including ADAM8, CTSD, CCL4, IL2RB, TTC38, and PLEK. Two validation cohorts from the GEO dataset, comprising 124 and 201 samples respectively, were used. The signature was significantly associated with overall survival and correlated with immune cell infiltration, immune and stromal scores, and immune checkpoint genes. Furthermore, the signature was associated with the homologous recombination deficiency (HRD) and T-cell receptor (TCR) scores. In conclusion, our study proposes a new prognostic signature based on NK cell marker genes, which may serve as a potential tool to predict overall survival and immunotherapy response for CRC patients.