Fig. 1: Study outline and selection process of tumor-supportive signature genes.

a Study approach. We first investigated the DEGs between tumors and NBT and designated overexpressed genes in tumors as tumor-supportive signature genes. We categorized the subgroups on the basis of the profile of tumor-supportive signature genes in NBTs (relatively high tumor-supportive signature scores in NBT as TSM; relatively low scores as HM). Our analysis incorporated multi-omics data—encompassing bulk RNA-seq data, H&E staining data of NBT, 16S rRNA-seq data and scRNA-seq data—and publicly available spatial transcriptomics data. b Summary of cohorts: bulk RNA-seq analysis was conducted on NBT and paired tumor samples from 273 patients. Among these, NBT and tumor samples from 47 patients were also subjected to scRNA-seq analysis. In addition, NBT and paired tumor samples from 44 patients were available for 16S rRNA-seq. The bulk RNA-seq data from TCGA, which included colorectal and other cancer types, were used for validation. c Volcano plot illustrating the DEGs between tumors and NBT. Red and blue dots represent overexpressed and underexpressed genes in tumors, respectively. Tumor-supportive signature genes are labeled. d Simulation results of the prognostic difference (Cox HR) between the TSM and HM groups. The x axis indicates the number of DEGs (mostly overexpressed in tumors) used for calculation, and the y axis indicates the HR (top) and −log P value (bottom) between the two groups (TSM versus HM). e Comparison of tumor-supportive signature scores by cohort characteristics with TNM stage, MSI status, medical center and tumor sidedness. f Comparison of tumor-supportive signature between NBT from the HM and TSM groups, colon polyp tissue, colon tissue from Crohn’s disease (both noninflamed and inflamed lesions) and tumor tissue from the HM and TSM groups. ^*Publicly available data for colon polyp tissue and Crohn’s disease (GSE208303). g CMSs according to the microenvironment-based classification (TSM and HM). (h) Clinical metadata of the study samples, including information on scRNA-seq analysis, 16S rRNA-seq analysis, disease recurrence, tumor location, CMSs, MSI status, TNM stage, adjuvant chemotherapy and medical center. i Comparison of the distance between paired NBT and tumor samples for the TSM and HM groups. The density plot shows the distribution of distances, while the scatter plot and box plot below further illustrate individual data points and summary statistics. DEGs, differentially expressed genes; NBT, non-tumor-bearing tissue; TSM, tumor-supportive microenvironment; HM, healthy microenvironment; H&E, hematoxylin and eosin staining; RNA-seq, RNA sequencing; scRNA-seq, single-cell RNA sequencing; 16S rRNA-seq, 16S rRNA sequencing; CRC, colorectal cancer; TCGA, The Cancer Genome Atlas; Rt, right-sided; Lt, left-sided; CMS, consensus molecular subtype; MSI, microsatellite instability; MSI-L, microsatellite instability–low; MSS, microsatellite stable; MSI-H, microsatellite instability–high; Adj chemo, adjuvant chemotherapy.