A predefined sulphur-handling transcriptional programme is associated with colorectal cancer molecular subtypes.

BACKGROUNDColorectal cancer (CRC) exhibits substantial metabolic and transcriptomic heterogeneity, but the relationship between sulphur-handling transcriptional programmes and consensus molecular subtypes (CMS) remains incompletely resolved.

METHODSWe analysed a predefined 14-gene sulphur-handling panel across The Cancer Genome Atlas colon and rectal adenocarcinoma cohorts (TCGA-COAD/READ) and six independent HG-U133 validation cohorts (n = 1,337 tumours). Analytical workflows integrated differential expression comparing tumours with adjacent-normal mucosa, CMS association modelling, random-effects meta-analysis, adjustment for three orthogonal tumour-purity estimators (ESTIMATE, consensus purity estimates, and ABSOLUTE), reference-based single-cell deconvolution with independent cross-atlas verification, and direction-aware therapeutic triage.

RESULTSColorectal tumours demonstrated branch-specific transcriptional alterations, marked by coordinated downregulation of sulphur oxidation and clearance genes (SQOR, ETHE1, TST) and of the cysteine-catabolism genes CDO1 and CTH, and selective upregulation of SLC7A11 and AHCY, whereas CBS was not significantly altered. The composite programme score differed across CMS subtypes and was consistently lower in mesenchymal CMS4 tumours than in non-CMS4 tumours across all discovery and validation series. This association remained statistically detectable after adjustment for three global tumour-purity metrics. The programme score showed no statistically detectable independent association with overall, recurrence-free, or disease-free survival. Single-cell deconvolution supported expression across epithelial, stromal, and immune compartments. Direction-aware therapeutic triage prioritised SLC7A11/xCT inhibition as an exploratory candidate for preclinical testing.

CONCLUSIONSulphur-handling transcriptional architecture is reproducibly linked to CRC molecular taxonomy and is not fully explained by measured tumour purity variation, providing a computational framework for future functional and compartment-resolved investigation of sulphur metabolism in stratified CRC populations.