Identification of long non-coding RNAs associated to the CMS molecular subtypes as predictive biomarkers of response to therapy in colorectal cancer.

SAULNIER D; ABBA MC; LACUNZA E

Encuentro; REUNION DE SOCIEDADES DE BIOCIENCIA; 2021

The consensus molecular subtypes (CMSs) of colorectal cancer (CRC) define tumor heterogeneity at the gene-expression level. The clinical utility of the CMS classification resides in the possibility to estimate survival (prognostic value) and select patients for both chemotherapy and currently used targeted agents (predictive value). Long Non-coding RNAs (lncRNA) have been largely associated with cancer constituting an essential approach for the search and identification of these biomarkers. Aim: To identify lncRNA signatures with prognostic and predictive value associated with each CMS subtype. Methods: We performed an integrative bioinformatics analysis on GDC-TCGA-CRC dataset (n=674), considering CMS and SFM based-classification that discriminate tumor microenvironment and drug sensitivity. We first classify tumors in CMSs. We annotated lncRNAs (n=14084) and applied DESeq2 for the comparison of each CMS (CMSk) versus the rest (CMSK-k; pvalue 1). The obtained lists of the up-modulated lncRNAs exclusive of each CMS were evaluated according to CRC molecular features; the immune, stromal or epithelial tumor component; and the prognostic and predictive value (p