Cargando…

A novel non-linear approach for establishing a QSAR model of a class of 2-Phenyl-3-(pyridin-2-yl) thiazolidin-4-one derivatives

In this investigation, we aimed to address the pressing challenge of treating osteosarcoma, a prevalent and difficult-to-treat form of cancer. To achieve this, we developed a quantitative structure-activity relationship (QSAR) model focused on a specific class of compounds called 2-Phenyl-3-(pyridin...

Descripción completa

Detalles Bibliográficos
Autores principales: Wu, Leilei, Chen, Yonglin, Duan, Kangying
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Frontiers Media S.A. 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10565811/
https://www.ncbi.nlm.nih.gov/pubmed/37829302
http://dx.doi.org/10.3389/fphar.2023.1263933
Descripción
Sumario:In this investigation, we aimed to address the pressing challenge of treating osteosarcoma, a prevalent and difficult-to-treat form of cancer. To achieve this, we developed a quantitative structure-activity relationship (QSAR) model focused on a specific class of compounds called 2-Phenyl-3-(pyridin-2-yl) thiazolidin-4-one derivatives. A set of 39 compounds was thoroughly examined, with 31 compounds assigned to the training set and 8 compounds allocated to the test set randomly. The goal was to predict the IC(50) value of these compounds accurately. To optimize the compounds and construct predictive models, we employed a heuristic method of the CODESSA program. In addition to a linear model using four carefully selected descriptors, we also developed a nonlinear model using the gene expression programming method. The heuristic method resulted in correlation coefficients (R (2)) of 0.603, 0.482, and 0.107 for R(2) (cv) and S(2), respectively. On the other hand, the gene expression programming method achieved higher R (2) and S(2) values of 0.839 and 0.037 in the training set, and 0.760 and 0.157 in the test set, respectively. Both methods demonstrated excellent predictive performance, but the gene expression programming method exhibited greater consistency with experimental values. The successful nonlinear model generated through gene expression programming shows promising potential for designing targeted drugs to combat osteosarcoma effectively. This approach offers a valuable tool for optimizing compound selection and guiding future drug discovery efforts in the battle against osteosarcoma.