First multi-target QSAR model for predicting the cytotoxicity of acrylic acid-based dental monomers

• Type: Article in International Scientific Journal
• Date: 2022-12-08

Authors

M Natalia DS Cordeiro (Editor) FCUP; António H.S. Delgado (Author) Other; Amit K. Halder (Author) FCUP

Journal

Title: Dental Materials

Vol. 38

Pages: 333-346

ISSN: 0109-5641

Publisher: Elsevier

Indexing

ISI Web of Knowledge - 3 Citations

Other information

Authenticus ID: P-00W-625

DOI: 10.1016/j.dental.2021.12.014

Abstract (EN): Objective: Acrylic acid derivatives are frequently used as dental monomers and their cy- totoxicity towards various cell lines is well documented. This study aims to probe the structural and physicochemical attributes responsible for higher toxicity of dental mono- mers, using quantitative structure-activity relationships (QSAR) modeling approaches. Methods: A regression-based linear single-target QSAR (st-QSAR) model was developed with a comparatively small dataset containing 39 compounds, the cytotoxicity of which has been assessed over the Hela S3 cell line. By contrast, a classification-based multi-target QSAR model was developed with 138 compounds, the cytotoxicity of which has been re- ported against 18 different cell lines. Both models were set up following rigorous validation protocols confirming their statistical significance and robustness. Results: The performance of the linear mt-QSAR model, developed with various feature selection and post-selection similarity searching-based schemes, superseded that of all non-linear models produced with six machine learning methods by hyperparameter opti- mization. The final derived st-QSAR and mt-QSAR linear models are shown to be highly predictive, as well as revealing the crucial structural and physicochemical factors re- sponsible for higher cytotoxicity of the dental monomers. Significance: This study is the first attempt on unveiling the cytotoxicity of dental mono- mers over several cell lines by means of a single multi-target QSAR model. Further, such a model is ready to get widespread applicability in the screening of new monomers, judging from its almost accurate predictions over diverse experimental assay conditions.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 14