Improved QSAR modeling of anti-HIV-1 acivities by means of the optimized correlation weights of local graph invariants

Abstract

AbstractWe report the results derived from the use of molecular descriptors calculated with the correlation weights (CWs) of local graph invariants for modeling of anti-HIV-1 potencies of two groups of reverse transcriptase inhibitors. The presence of different chemical elements in the molecular structure of the inhibitors and the Morgan extended connectivity values of zeroth-, first-, and second order have been examined as local graph invariants in the labeled hydrogen-filled graphs. We have computed via Monte Carlo optimization procedure the values of CWs which produce the largest possible correlation coefficient between the numerical data on the anti-HIV-1 potencies and those values of the descriptors on the training set. The model of the anti-HIV-1 activity obtained with compounds of training set by means of optimization of correlation weights of chemical elements present together with Morgan extended connectivity of first order makes up a sensible model for a satisfactory prediction of the endpoints of the compounds belonging to the test set.