Abstract: This paper presents a predictive model for the retention times of aroma compounds in Kuerle fragrant pear wine on gas chromatography-mass spectrometry (GC-MS) based on quantitative structure-retention relationship (QSRR) analysis. Some of the molecular connectivity indexes and Kappa shape indexes of 54 aroma compounds in Kuerle fragrant wine were calculated by program according to molecular bonding valence state and topological environment, and were related with the retention times of these compounds on GC-MS through multiple linear regression based on the molecular topological theory to obtain the regression equation describing retention time/molecular connectivity and Kappa shape indexes relationship, which exhibited a correlation coefficient of 0.952. The robustness and predictive ability of the regression model was validated by jackknife method, and the results showed that the correlation coefficient was between 0.946 and 0.956.

Key words: topological index, Kuerle fragrant pear wine, retention time, aroma compounds, quantitative structureretention relationship (QSRR)