Abstract
The lack of target therapies is accountable for the higher mortality of various types of cancer. To address this issue, we selected a target mutated Kirsten rat sarcoma virus oncogene homologue, which plays a significant role in various cancers. Our study aims to identify selective biomarkers and develop diagnostic and therapeutic strategies for KRAS-associated genes using artificial intelligence. Initially, Genomic data, cancer epidemiology, proteomics network interactions, and omics enrichment were analyzed. Structured E-pharmacophore model aided in capturing the binding cavity using eraser algorithms and fabricating a new selective lead compound for the KRSA. The selective molecule was abridged inside the binding cavity and stability was validated through 100 ns molecular dynamics simulations. Epidemiological-neural network studies indicated KRAS mutations leads 40 types of cancer, exclusively pancreatic and colorectal cancers, with diploid and missense mutations as primary factors. Pathway analysis highlighted the involvement of the MAPK and RAS signaling pathways in cancer development and proteomics analysis identified RALGDS as a key protein. Protein-based pharmacophore analysis mapped the biologically active features such as donor, acceptor and aromatic ring with the designed ligands. The results of interaction interpretation illustrate that the amino acid Tyr566 formed an H-bond interaction with the amine group of the octyl ring system and 20 amino acids crafted to properly orient the molecule to fit inside the polar cavity of KRAS protein. The MMGBSA score of - 53.33 kcal/mol conformed to the well-configured binding with KRSA and realistic model simulation exposed the π-π, π-cationic and hydrophobic interactions stabilised the molecule inside the KRSA protein throughout 100 ns simulation. The study demonstrates the vitality of AI and network pharmacology to identify potential-target biomarkers for KRAS-associated genes, paving the way for improved cancer diagnostics and therapeutics.