Application of Artificial Intelligence for Predicting of New Potential Inhibitors of Vitamin K Epoxide Reductase

Artificial intelligence (AI) integration in drug development has altered the pharmaceutical industry by delivering innovative and effective solutions to long-standing difficulties in this field. AI techniques and tools, such as machine learning and deep learning algorithms, can analyze enormous datasets and identify complex patterns, therefore speeding up the drug discovery process. The research presented in this paper introduces a systematic approach to the design of novel anticoagulative agents with significant therapeutic potential, driven by the utilization of AI in drug design and development. Coumarins are compounds isolated from plants, and they are responsible for a wide range of biological functions. The anticoagulant potential of coumarin derivatives is one of their most prominent pharmacological properties. Warfarin (WFR), a common oral anticoagulant, belongs to the 4-hydroxycoumarin class. In one of our most recent studies, we performed pharmacological profiling and anticoagulant activity testing of a (E)-3-(1-((4-hydroxy-3-methoxyphenyl)amino)ethylidene)-2,4-dioxochroman-7-yl acetate [15]. When compared to the WFR, it was discovered that the examined molecule has better anticoagulative capability, as well as a more desirable pharmacokinetic profile than WFR. Furthermore, preliminary laboratory tests revealed that (E)-3-(1-((4-hydroxy-3-methoxyphenyl)amino)-ethylidene)chromane-2,4-dione (L) has similar pharmacokinetic properties with increased water solubility, making it an even better candidate for the development of potential anticoagulants. By utilization of the CReM web server, a series of 1000 derivatives of L were generated and narrowed down to 46 compounds through screening based on drug-likeness rules. Toxicity assessments further refined the selection to 16 non-toxic candidates, subsequently subjected to molecular docking simulations. Four compounds emerged from this subset, demonstrating excellent inhibition of VKOR, and VKORC1, compared to the conventional anticoagulant WFR. Additional investigation through molecular dynamics simulations provided insights into their activity in a specific timeframe confirming the results obtained from molecular docking simulations. The obtained results strongly suggest that investigated compounds have shown promising anticoagulative potential and should be further tested.