Abstract: In this article, a novel connectivity index (mLtv) was established by extending Kier,s connectivity index (mXtv). The quantitative structure-activity relationship (QSSR) between the relative sweetness (RS/B) of 20nitroaniline and cyanoaniline derivatives and mLtv was explored using multivariate statistical regression. Based on leaps-and-bounds regression analysis, an optimal binary QSAR model was set up. The traditional correlation coefficient (R) and cross-validation correlation coefficient (Q2) of leave-one-out (LOO) were 0.943 and 0.844, respectively. These results demonstrated that the model is highly reliable and has good prediction capability. Meanwhile, it was better than that of Kier,s connectivity index. Moreover, the model can be explained by the AH-B-W sweetness theory of Shallenberger and Kier.

Key words: nitroaniline, cyanoaniline, relative sweetness, molecular connectivity index, quantitative structure-activity relationship