Format

Send to

Choose Destination
Int J Mol Sci. 2019 Jun 8;20(11). pii: E2801. doi: 10.3390/ijms20112801.

Quantitative Structure-Activity Relationships for Structurally Diverse Chemotypes Having Anti-Trypanosoma cruzi Activity.

Author information

1
Laboratory of Computational and Medicinal Chemistry, Center for Research and Innovation in Biodiversity and Drug Discovery, Physics Institute of Sao Carlos, University of Sao Paulo, Sao Carlos-SP 13563-120, Brazil. anacletosilvadesouza@usp.br.
2
Laboratory of Computational and Medicinal Chemistry, Center for Research and Innovation in Biodiversity and Drug Discovery, Physics Institute of Sao Carlos, University of Sao Paulo, Sao Carlos-SP 13563-120, Brazil. leonardo@ifsc.usp.br.
3
Laboratory of Computational and Medicinal Chemistry, Center for Research and Innovation in Biodiversity and Drug Discovery, Physics Institute of Sao Carlos, University of Sao Paulo, Sao Carlos-SP 13563-120, Brazil. aldo.sena@ufsc.br.
4
Department of Exact Sciences and Education, Blumenal Center, Federal University of Santa Catarina, Blumenau 89036-256, Brazil. aldo.sena@ufsc.br.
5
Laboratory of Computational and Medicinal Chemistry, Center for Research and Innovation in Biodiversity and Drug Discovery, Physics Institute of Sao Carlos, University of Sao Paulo, Sao Carlos-SP 13563-120, Brazil. aandrico@ifsc.usp.br.

Abstract

Small-molecule compounds that have promising activity against macromolecular targets from Trypanosoma cruzi occasionally fail when tested in whole-cell phenotypic assays. This outcome can be attributed to many factors, including inadequate physicochemical and pharmacokinetic properties. Unsuitable physicochemical profiles usually result in molecules with a poor ability to cross cell membranes. Quantitative structure-activity relationship (QSAR) analysis is a valuable approach to the investigation of how physicochemical characteristics affect biological activity. In this study, artificial neural networks (ANNs) and kernel-based partial least squares regression (KPLS) were developed using anti-T. cruzi activity data for broadly diverse chemotypes. The models exhibited a good predictive ability for the test set compounds, yielding q2 values of 0.81 and 0.84 for the ANN and KPLS models, respectively. The results of this investigation highlighted privileged molecular scaffolds and the optimum physicochemical space associated with high anti-T. cruzi activity, which provided important guidelines for the design of novel trypanocidal agents having drug-like properties.

KEYWORDS:

Chagas’ disease; QSAR; artificial neural networks; machine learning; molecular modeling

PMID:
31181717
DOI:
10.3390/ijms20112801
Free full text

Supplemental Content

Full text links

Icon for Multidisciplinary Digital Publishing Institute (MDPI)
Loading ...
Support Center