Objective:  Colorectal cancer (CRC) is one of the leading causes of cancer-related mortality in the world. Incidence and mortality are predicted to rise globally during the next several decades. While colon cancer-related deaths are quite common in wealthy countries, the disease's prevalence and fatality rate are also rising in developing nations. Recent advancements in research have highlighted the significant role of microbes and proteins in the progression of colon cancer. This study delineates the interactions between microbes and proteins, focusing on critical signaling pathways implicated in CRC progression.

Methods: We conducted comprehensive proteomics analysis using two datasets, one from the UK Biobank (Training dataset, N=518) another from the Bosch et.al. (Validation dataset, N=33). For biomarker discovery we used multiple machine learning methods like XGBoost, LASSO, LightGBM along with explainable artificial intelligence to identify top proteins. Further, we used functional pathway analysis to identify protein -protein interaction and Human protein atlas. In addition we linked protein biomarkers with the microbiome diversity in the pathways involved in the colon cancer progression.

Results: A total of 159 proteins were found to be significant in UK Biobank and used to the three models and resulted in 7 significant proteins (LCN2, TFF1, TFF3, AHCY, CEACAM5, SELE, RETN) which was also upregulated in Human protein atlas. Seven proteins were found to be involved in five pathways: PI3K/Akt, ERK signaling, JAK-STAT, DNA methylation, Caspase activation. We further found Fusobacterium nucleatum, Bacteroides fragilis, Escherichia coli, Enterococcus faecalis, and Helicobacter pylori to have a potential interaction with those pathways.