Generic QSPR study for predicting critical micelle concentration of gemini cationic surfactants using the online chemical modeling environment (OCHEM)

Abstract

In this paper, a generic QSPR (quantitative structure-property relationship) model have been developed to investigate the relation between the critical micelle concentration (cmc) and the molecular structure of 231 gemini cationic surfactants with the various hydrophilic head group. The QSPR modeling were performed using the OCHEM (Online CHEmical Modeling environment), a web-based model development platform. The OCHEM platform provides several molecular descriptors calculation and machine learning methods as a tool to build QSPR models. Eight different software packages including Dragon v6, OEstate and ALogPS, CDK, ISIDA Fragment, ChemAxon, Inductive Descriptor, Mordred, and PyDescriptor are used to calculate molecular parameters of gemini cationic surfactants. Also, eight machine learning methods (MLRA, ASNN, kNN, LibSVM, FSMLR, DNN, RFR, and PLS) are used to develop QSPR models. A total of 64 QSPR models were generated with 6 top-ranked models. Based on the statistical coefficient of QSPR models, the model 5 which is constructed from combination of ASNN method and Mordred descriptors, provided the best QSPR models. The model 5 performed the highest predictive result with R2 = 0.95, q2 = 0.95, RMSE = 0.17, and MAE = 0.11. The model can be access on OCHEM website (https://ochem.eu/model/25147470) and can be used for prediction of cmc of new gemini cationic surfactants compound at the early steps of gemini cationic surfactants development.