Artificial intelligence and in silico study using De Novo drug design to explore molecular targets of Burkholderia pseudomallei for melioidosis treatment

Melioidosis is an infectious disease caused by the bacterium Burkholderia pseudomallei . Transmission occurs through contact with contaminated soils and waters, and there is a higher risk for individuals with compromised immune systems. This bacterium exhibits intrinsic resistance to multiple antibiotics, which complicates its diagnosis, particularly in non-endemic regions where selective diagnostic methods are absent. This study sought to integrate an artificial intelligence module (MolAICal) into the discovery of novel inhibitors of capsular polysaccharides (CPSs), associated with bacterial defense. Molecular docking simulations were performed using the AutoDockVina code, while molecular dynamics (MD) simulations were performed using the NAMD software. The pharmacokinetic prediction of the lead compounds was performed using ADMETlab 3.0, ADMET-AI, and admetSAR 3.0 tools, while metabolic stability was estimated using the XenoSite, STopTox, and GUSAR Online servers. Among 842 hits, MolAICal filtered Hit 121 by the alignment between affinity energy (–9.74 kcal/mol) and synthetic accessibility (> 60%), which interacted with the charged portion of residue Arg 167, in the CoAF binding domain of the CPSs. MD simulations indicated a binding free energy (ΔG) of around –18.0 kcal/mol. Furthermore, pharmacokinetic viability is estimated by a high cell permeability (in the order of 10 -6 cm/s) and microsomal clearance around 50 µL/min/mg, predicting good oral bioavailability. Despite the therapeutic viability estimated by the simulations, molecular tests are essential and constitute future steps within the scope of this work.