Cyanobacterial secondary metabolites: Genotoxicity and carcinogenicity in silico evaluation

Cyanobacterial secondary metabolites are considered to have potential as biologically active compounds (e.g., antibacterial, antifungal) and are attracting great interest in medical applications and industrial innovation. However, it is well established that a large proportion of newly identified molecules, including natural compounds, are ultimately rejected due to adverse effects on human health. In silico tools could be a rapid and efficient way to screen the toxicity properties of secondary metabolites with potential biotechnological applications. In this study, 25 cyanobacterial secondary metabolites with known bioactivity were selected from CyanoMetDB (Cyanobacterial Secondary Metabolites Database) - a comprehensive and publicly available database - and screened for the toxicological endpoints genotoxicity and carcinogenicity using different in silico models and/or tools (VEGA, OECD QSAR Toolbox, T.E.S.T. (Toxicity Estimation Software Tool), ADMETlab, Carcinogenicity Predictor). A majority-consensus approach was applied to integrate the predictions. Seven secondary metabolites were predicted as non-genotoxic and non-carcinogenic, suggesting they could be promising candidates for biocidal and/or other biotechnological applications. The results support integrating in silico approaches for screening and prioritizing cyanobacterial secondary metabolites, while highlighting limitations related to the applicability domain and the underrepresentation of complex natural products in model training sets. In the future, NAMs-based data generated for microbial secondary metabolites could improve the representation of these compounds in the in silico model datasets, thereby increasing the robustness of the predictions.