Practical Guidelines for Optimising Free Energy Calculations Using Thermodynamic Integration

Free energy calculations have become valuable in structure-based drug design, but their accuracy and reliability remain significant challenges. Here, we present an automated workflow for free energy estimation built with AMBER20, alchemlyb, and an open-source cycle closure algorithm. Evaluated on 178 perturbations across four datasets, the short sub-nanosecond simulations performed comparably or better than prior studies for the MCL1, BACE, and CDK2 datasets, while the TYK2 dataset required a longer equilibration time (∼ 2 ns). Perturbations with |ΔΔG| > 2.0 kcal/mol exhibited higher errors, suggesting such perturbations are unreliable, hence providing a practical guideline for improving thermodynamic integration simulations.