Year: 2026
Type: article
Abstract: In the present report 29 ( E )-stilbenes are clustered by using a procedure based on artificial intelligence. The objective is to predict cytotoxicity (anticancer activities) of them and other similar stilbenes in nine cell lines. We make a periodic classification of stilbenes by using the information entropy theory to select the most active classes. The structures of seven different classes are obtained. The most active Class 1 is located at the bottom right of the periodic classification. We provide new compounds that would also have high activity because they are in the bottom-right groups. Moreover, we relate stilbenes’ cytotoxicity in the cell lines to their physical and chemical properties by QSAR and PCA. The scores plot separates clusters that contain classes in the periodic system. The results of the nine QSAR models are good with r 2 greater than 0.57 and show the repetition of most variables for all the nine cell lines. The leave- m -out cross-validation determines the good robustness for the cell lines ( q 2 > 0.33). These results agree with the loading plot and suggest the importance of these descriptors in further studies of other cell lines.
Source: Chemical Papers
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,490
Funder: Universitat de València
OpenAlex ID: W7164892900