New Methods for Pathogen Risk Assessment: Machine Learning in the Analysis of Toxicity Spectrum of <i>Albifimbria verrucaria</i>

Abstract

HighlightsThe use of artificial intelligence has great potential for predicting the toxic properties of new little-studied chemical compounds, reducing the time and financial costs associated with identifying the risks of possible threats.Relevance. Mycotoxins, which are secondary metabolites of mold fungi, represent one of the most significant factors of chronic risk associated with food products. Their danger exceeds the threat posed by synthetic pollutants, plant toxins, food additives, and pesticide residues. However, for many mycotoxins, the full toxicological profile has not yet been established, and traditional analysis methods remain labor-intensive, costly, and insufficiently effective. This makes the search for new approaches to assess their danger and control highly relevant.Purpose of the study is to study the toxicological profile of mycotoxins produced by the pathogenic fungus Albifimbria verrucaria and to determine their level of danger using chemoinformatics and machine learning.Study base sources. Analysis of scientific literature available through open Russian and English-language Internet resources.Method. In silico methods were applied to analyze the toxicological profile of mycotoxins, enabling the identification of high-risk compounds. These methods prioritize substances for further in-depth toxicological assessment, significantly reducing the time and resources required for research.Results and Discussion. The study results showed that approximately 50% of mycotoxins produced by mold fungi belong to hazard classes I and II. At the same time, a significant portion of these compounds remains outside the control zone, despite their potential threat to living organisms. This highlights the need for more thorough study and monitoring of such substances.Conclusions. The obtained data confirm the importance of developing and implementing modern systems for monitoring and regulating mycotoxins, especially for poorly studied and new compounds. The use of chemoinformatic methods makes it possible to effectively identify the most hazardous substances and focus efforts on their research, thereby enhancing food safety and reducing risks to human and animal health.

Keywords

• aflotoxin
• albifimbria verrucaria
• chemoinformatics
• in silico toxicity assessment
• machine learning
• mycotoxins
• myrothecium verrucaria
• predictive toxicology
• roridin
• trichothecene mycotoxin
• verrucarin