Vladimir Poroikov » Публикация

Аннотация In the existing QSAR methods any molecule is represented as a single point in many-dimensional space of molecular descriptors. We proposed a new QSAR approach based on Quantitative Neighbourhoods of Atoms (QNA) descriptors, which characterize each atom of a molecule and depend on the whole molecule structure. In “Star Track” methodology any molecule is represented as a set of points in two-dimensional space of QNA descriptors. By our new method the estimate of target property of chemical compound is calculated as the average value of the function of QNA descriptors in the points of the atoms of a molecule in QNA descriptors’ space. Substantially, we have proposed to use only two instead of more than three thousand molecular descriptors applying in QSAR. On the basis of this approach we have developed computer program GUSAR and compared it with the several widely used QSAR methods including CoMFA, CoMSIA, Golpe/GRID, HQSAR and others, using ten data sets representing various chemical series and diverse types of biological activity. It was shown that in the majority of cases the accuracy and predictivity of GUSAR models appeared to be better than for the reference QSAR methods. High predictive ability and robustness of GUSAR were also shown in leave-20%-out cross-validation procedure.