Abstract:
Colorectal cancer (CRC) remains a leading cause of cancer-related deaths, with
chemoresistance significantly hindering treatment efficacy. This study explores the
influence of the gut microbiome on chemoresistance in CRC patients, aiming to identify
microbial genes and species associated with treatment outcomes. The study analyzed
metagenomic data from CRC patients categorized into groups with progressive disease and
partial response. Validation was conducted using a separate dataset with untreated and
treated CRC samples. In this analysis key microbial genes were identified, such as butyrate,
ABC family, deoxy uridine triphosphatase, histone deacetylase, glutathione s-transferase,
cytidine deaminase, cystine desulfurase, ribonucleotide reductase, and thymidylate
synthase, which were significantly associated with chemoresistance. The high abundance
of these genes correlated with poor treatment responses, suggesting their roles in
modulating drug efficacy. Taxa association analysis revealed species Akkermansia
muciniphila, Escherichia coli, Ruminococcus albus, Clostridium saccharolyticum, and
Ethanoligenens harbinense to be prevalent in non-responders, indicating their potential
involvement in CRC progression and therapeutic response. Machine learning models,
particularly Gradient Boosting, demonstrated high accuracy in classifying
chemoresistance-associated genes and species. These findings provide novel insights into
the microbial mechanisms underlying chemoresistance in CRC, identifying potential
biomarkers and therapeutic targets to enhance treatment efficacy. This study highlights the
critical role of the gut microbiome in CRC treatment response, offering avenues for
xv
developing microbiome-based interventions to overcome chemoresistance and improve
patient outcomes.
