MLDockKit: A precision drug discovery platform for prostate cancer

Prostate cancer (PC) presents a significant challenge in men's health, demanding therapeutic options beyond Androgen deprivation therapy (ADT). PC is a type of cancer that forms in the prostate gland. The prostate is a small, walnut-shaped gland responsible for producing seminal fluid that nourishes and transports the sperm. Prostate cancer is a slow growing disease common in men above 50 years and is the leading cause of cancer deaths after lung cancer presenting a significant challenge in men's health. Recent research has highlighted the role of estrogen receptor alpha (ERα) in PC progression, offering a promising avenue for targeted intervention. However, the protracted drug discovery process necessitates innovative solutions. This study introduces MLDockKit, a Python toolkit designed for cancer therapeutics, particularly in prostate cancer management. This tool trained on a curated dataset from the ChEMBL database, MLDockKit employs PyCaret package for regression modeling, with the random forest regressor demonstrating superior performance within short period. Having a 74 % prediction accuracy MLDockKit showed significant similarities between its’ predicted pIC50(logarithmic Half maximal inhibitory concentration) and ChEMBL experimental pIC50. MLDockKit accelerates the search for ERα-targeting compounds through Lipinski descriptor computation; fundamental guidelines in the assessment of compounds drug likeliness, prediction of compound pIC50 values; a measure of the potency of an antagonist drug in pharmacological research, and molecular docking simulations of the selected ligand (molecule) and the target protein (ERα). This capability facilitates the exploration of binding interactions between the two, culminating in intuitive visualization of docking results. MLDockKit docking score of −9.111 kcal/mol is comparable to other docking tools such AutoDock Vina, Smina and LeDock in addition to offering full automation of ligand preprocessing, docking box definition, and visualization. MLDockKit aims to empower researchers in identifying potential drug candidates targeting ERα for PC management, streamlining the drug discovery process efficiently.