2.1. Data collection and identification of MRG

Using the TCGA (https://portal.gdc.cancer.gov/) and GEO (https://www.ncbi.nlm.nih.gov/geo/) databases, researchers were able to access transcriptome data, clinical data, and mutation data from CRC samples and normal tissue samples. The TCGA cohorts includes TCGA‐COAD (41 normal tissue samples, 480 tumor tissue samples) and TCGA‐READ (6 normal tissue samples, 161 tumor tissue samples). The GEO cohorts contains {"type":"entrez-geo","attrs":{"text":"GSE17538","term_id":"17538"}}GSE17538 (238 samples) ³⁸ and {"type":"entrez-geo","attrs":{"text":"GSE39582","term_id":"39582"}}GSE39582 (585 samples). ³⁹ In the process of organizing clinical data, samples without a survival time and samples with a survival time of 0 are excluded. The collection of MRGs are organized through the MITOMAP (https://www.mitomap.org/MITOMAP). By removing duplicate genes, we can obtain an overall of 1592 MRGs.