Obtaining quantitative structure-activity relationship (QSAR) models for prediction to antibacterial activity on heterogeneous series of compounds

Authors

  • Argenis Soutelo-Jiménez
  • América García-López
  • Julio Rojas-Vargas
  • Yennys Hernández-Molina

Abstract

Two discriminant models for prediction of antibacterial activity were obtained. Model 1 was obtained using descriptors TOPS-MODE and fragments, model 2 with 3D descriptors and fragments using Linear Discriminant Analysis. The study was performed with 402 compounds reported in the literature. Model 1 ranked 90 and 90 % of active cases and 97 and 93 % of inactive cases in training sets and prediction respectively, with an overall rating of 93 and 91 %. Model 2 ranked 89 and 90% of active cases and 95 and 91 % of inactive cases in training sets and prediction respectively, with an overall rating of 92 and 89 %. These results and the values of the statistical indices of the models allowed to show their qualities. In addition, the contributions were calculated fragments of antibacterial activity for both models.

 

Keywords: QSAR, antibacterial, TOPS-MODE, 3D descriptors, LDA.

Published

2016-05-19

How to Cite

Soutelo-Jiménez, A., García-López, A., Rojas-Vargas, J., & Hernández-Molina, Y. (2016). Obtaining quantitative structure-activity relationship (QSAR) models for prediction to antibacterial activity on heterogeneous series of compounds. Revista Cubana De Química, 28(1), 462–489. Retrieved from https://cubanaquimica.uo.edu.cu/index.php/cq/article/view/697

Issue

Section

Artículos

Most read articles by the same author(s)