In Silico Prediction of New Inhibitors for Kirsten Rat Sarcoma G12D Cancer Drug Target Using Machine Learning-Based Virtual Screening, Molecular Docking, and Molecular Dynamic Simulation Approaches

Abstract

 The majority of human cancers are caused by mutations in the Kirsten rat sarcoma (KRAS) 
viral proto-oncogene. Approximately 30% of malignancies in the human body originate from 
oncogenic KRAS mutations, which mostly affect the colon, pancreas, and lungs. Pancreatic cancer is 
caused by one of the most common mutant KRAS G12D forms, which makes it an appealing 
therapeutic target. Currently, there are no medications that have been authorized for the KRAS G12D 
mutant by the Food and Drug Administration (FDA). The development of a potent medication for 
KRAS G12D is therefore imperative. It takes a lot of effort and money to discover new medications. 
Additionally, in silico drug development approaches save time and money. In order to find novel 
inhibitors for the KRAS G12D mutant, we used machine learning methods including K-nearest 
neighbor (KNN), support vector machine (SVM), and random forest (RF). It was anticipated that 82 
hits would be active against the KRAS G12D mutant. Docking the active hits into the KRAS G12D 
mutant's active site was the process. Also, the top two complexes and the reference complex (MRTX
1133) were simulated using MD for 200 ns to assess the stability of the compounds with strong docking 
scores. As compared to the conventional compound, the top two hits demonstrated great stability. Both 
of the top finds had higher binding energies than the reference compound. We have uncovered hits 
that may help fight cancer by inhibiting the KRAS G12D mutation. Our research indicates that this is 
the first study to date to use molecular dynamics modeling, molecular docking, and virtual screening 
based on machine learning to find potential inhibitors for the KRAS G12D mutant.