Quantitative Structure Property Relationship Modeling for Prediction of Retention Index for a Set of Some Organic Compounds

One of the most ubiquitous challenges of the scientists is the theoretical evaluation of experimental parameters to validate and improve their ability. Plant essential oils and their extracts have been greatly employed in folk medicine, food flavoring, fragrance and pharmaceutical industries. This work is a part of our comprehensive investigation to correlate the experimental and calculated retention indices (RI) of the some organic compounds from K. Javidnia et al. The structures of all organic compounds were drawn into the HYPERCHEM program and optimized using semi-empirical AM1 method, applying a gradient limit of 0.01 kcal/ as a stopping criterion for optimized structures prior to geometry optimization step. Then molecular descriptors were calculated for each compound by the DRAGON software on the minimal energy conformations. The Stepwise SPSS was used for the selection of the variables that resulted in the best fitted models. By molecular modeling and calculation of descriptors, four significant descriptors (XMOD, PCD, MATS2e, GATS2e) related to the retention indices values of the essential oils, were identified. After the variables selection, the MLR method used for building the regression models. The statistical figures obtained by the proposed model are R2=0.989, RMSEP=53.08, REP =3.83 and SEP =54.94. In the final step, models generated were used to predict the retention index for a set of test compounds.