A new in silico classification model for ready biodegradability, based on molecular fragments

Area of Science:

• Environmental Chemistry
• Computational Toxicology

Background:

• Chemical safety regulations necessitate biodegradability assessments.
• In silico models offer efficient alternatives to experimental testing for biodegradability.

Purpose of the Study:

• To develop a novel in silico model for predicting chemical ready biodegradability.
• To provide a cost-effective and rapid screening tool for environmental risk assessment.

Main Methods:

• Utilized a dataset of 728 chemicals with ready biodegradability data (MITI-test).
• Employed SARpy software for automated extraction of statistically relevant structural fragments.
• Integrated fragments with expert-based rules to create a predictive model.

Main Results:

• Achieved high accuracy: 92% (training), 82% (test), and 76% (external set).
• Model performance is comparable to established tools like the US EPA's BIOWIN.
• The model provides easily understandable explanations for its predictions.

Conclusions:

• The developed in silico model is a reliable tool for predicting chemical ready biodegradability.
• This model can support regulatory compliance and reduce the need for extensive experimental testing.
• The model's interpretability enhances its practical application in risk assessment.