Colorectal cancer (CRC) is ranked second worldwide in cancer-related deaths. Despite many advancements in targeted therapies, the five year survival rate for CRC is largely discouraging because of the tumor diversity as well as the two-way communication pathways between the cancer cells and their immunosuppressive microenvironments. This study used machine-learning models to identify five genes that work together in defining a post-translational modification (PTM) subtype, highlighting the subtype as a useful mechanism for honing immunotherapy treatments. Overall, a combination of multi-omics analysis, transcriptomic data, mendelian randomization, differential expression, GSVA, and machine-learning were used to conduct the study. Shapley Additive Explanations (SHAP) analysis was implemented to analyze the functionalities of crucial genes in colorectal carcinogenesis before statistical analysis was conducted. The multi-omics data indicated that patients with colorectal cancer presented with dysregulation in 80% of their PTM pathways. The machine learning model identified a five gene model, introducing the possibility of distinguishing between patients of CRC and those without. Within these five genes, the gene GALNT6 was identified as a causal risk-factor of CRC.