A computationally affordable approach for accurate prediction of the binding affinity of JAK2 inhibitors

Janus kinase 2 (JAK2) inhibitors are potential anticancer drugs in the treatment of lymphoma, leukemia, thrombocytosis and particularly myeloproliferative diseases. However, the resemblance among JAK family members has challenged the identification of highly selective inhibitors for JAK2 to reduce undesired side effects. As a result, a robust search for promising JAK2 inhibitors using a computational approach that can effectively nominate new potential candidates to be further analyzed through laborious experimental operations has become necessary. In this study, the binding affinities of JAK2 inhibitors were rapidly and precisely estimated using the fast pulling of ligand (FPL) simulations combined with a modified linear interaction energy (LIE) method. The approach correlates with the experimental binding affinities of JAK2 inhibitors with a correlation coefficient of R = 0.82 and a root-mean-square error of 0.67 kcal•mol^-1. The data reveal that the FPL/LIE method is highly approximate in anticipating the relative binding free energies of known JAK2 inhibitors with an affordable consumption of computational resources, and thus, it is very promising to be applied in in silico screening for new potential JAK2 inhibitors from a large number of molecules available.