BACKGROUND: Colon cancer and pancreatic ductal adenocarcinoma are among the most aggressive tumors for which therapeutic options are limited. Both cancers share common features, such as some KRAS pathogenic variants and common epidemiology. The integration of multidimensional datasets by combining machine learning and bioinformatics approaches could provide deeper insights into the intricate KRAS-related networks underlying cancer progression and unveil novel biomarkers and potential therapeutic targets. This study aimed to uncover colon and pancreatic cancers that shared transcriptional changes closely related to KRAS missense mutations. METHODS: Feature Selection (FS) technique and Qiagen's Ingenuity Pathway Analysis (IPA) were used to combine DNA-Seq and RNA-Seq data from mutant and wild-type (WT) KRAS colon and pancreatic tumor samples. RESULTS: From the FS, we prioritized 70 genes (54 protein-coding genes and 16 ncRNA-coding genes) that were able to discriminate between WT and mutated KRAS patients. These genes were involved in KRAS signaling and other related processes, such as EMT signaling, glycolysis, apical junction, Wnt/beta-catenin signaling, and IL-2/STAT5 signaling. Using IPA, we identified a top-scoring network of 19 upregulated genes in both tumor types stratified into mutant KRAS and WT KRAS samples. For a set of genes, qRT-PCR performed on colon and pancreatic representative cancer cell lines showed concordant expression trends when comparing colon-dominant KRAS mutants versus WT KRAS and dominant pancreatic KRAS mutants versus WT KRAS, as expected according to in silico analyses. CONCLUSIONS: Our findings may provide insight into the common transcriptional signatures potentially underlying colon and pancreatic KRAS-mutant cancers. However, further studies are needed to elucidate the diagnostic and prognostic value of targets identified as common features in our study.