The Model for Prediction of Inhibition of Influenza A and B Neuraminidases Based on a Reduced Set of Energy Terms
A.V. Mikurova, Vladlen S. Skvortsov · Biochemistry (Moscow) Supplement Series B Biomedical Chemistry · 2020
Abstract— A general model for predicting the IC50 value (concentration causing 50% inhibition of enzyme activity) was obtained. It combined data for 40 variants of neuraminidases of influenza virus A (7 serotypes) and B and three known neuraminidase inhibitors, oseltamivir, zanamivir, and peramivir and using data only on enthalpy contributions to the potential energy of the inhibitor/protein and substrate/protein complexes (MUNANA). The calculations have been adapted to the use software with support for graphics accelerators, which reduces the calculation time. The R2 values for predicting pIC50 values for test sets were in the range of 0.45−0.58, the average error was 0.7 (with the range of the pIC50 training set from 4.55 to 10.22).