Alignment In dependent ۳D-QSA R Modeli ng of Full erene (C۶۰) Solubility in Different Organic Solvents

In this work, we have app lied VolSurf and GRIND derived descriptors of Mol ecular Interaction Fields (MIFs), as Alignment independent three dimensional quantitative stru cture activity relationships (۳D-QS AR) approac hes, to pred iction of C۶۰ solubility in a diverse set of ۱۳۲ organic solve nts. GRIN D methodol ogy with fra ctional factorial design (FFD) and PLS analysis was carr ied out an d yielded a highly des criptive an d predictive model. G enetic algorithm (GA) and successive projection algorithm (SPA) applied to feature selection and extract more informative Vol Surf descri ptors. Supp ort vector m achine (SVM) were used to model construction and SPA- SVM based VolSurf d escriptors sh owed excellent perfor mance in predicting C۶۰ solubility. V alidation, r eliability an d robustness of obtained models checked by the prediction of external test sets, leav e-one-out and progressi ve scrambling approach . The results confirmed that hydrophobic in teractions b esides steric effects are main f actors influencing solubility of C۶۰ i n different organic solvents.

کلیدواژه ها:

Fullerene (C۶۰) ، Solubility ، ۳ -QSAR ، G RIND ، VolSurf