Quantitative Structure-Retention Relationships Study of Phenols Using Neural Network and Classic Multivariate Analysis

Abstract

A quantitative structure-retention relationship (QSRR) study, has been carried out on 50 diverse phenols in gas chromatography (GC) in a dual-capillary column system made of DB-5 (SE-54 bonded phase) and DB-17 (OV-17 bonded phase) fused-silica capillary columns by using molecular structural descriptors. Modeling of retention times of these compounds as a function of the theoretically derived descriptors was established by multiple linear regression (MLR), partial least squares (PLS) regression and artificial neural networks (ANN). Stepwise SPSS was used for the selection of the variables (descriptors) that resulted in the best-fitted models. For prediction retention times of compounds in DB-5 and DB-17 columns, three and four descriptors, respectively were used to develop a quantitative relationship between the retention times and structural properties. Appropriate models with low standard errors and high correlation coefficients were obtained. After variables selection, compounds randomly were divided into two training and test sets and MLR and PLS methods (with leave-one-out cross validation) and ANN used for building of the best models. The predictive quality of the QSRR models were tested for an external prediction set of 10 compounds randomly chosen from 50 compounds. The squared regression coefficients of prediction for the MLR, PLS and ANN models for DB-5 column were 0.9645, 0.9606 and 0.9808, respectively and also for DB-17 column were 0.9757, 0.9757 and 0.9875, respectively. Result obtained showed that non-linear model can simulate the relationship between structural descriptors and the retention times of the molecules in data sets accurately.