Prediction of atmospheric degradation data for POPs by gene expression programming

Area of Science:

• Environmental Chemistry
• Computational Chemistry
• Cheminformatics

Background:

• Persistent organic pollutants (POPs) pose significant environmental risks due to their slow degradation.
• Accurate prediction of atmospheric degradation half-lives is crucial for assessing POPs' environmental fate and impact.
• Quantitative structure-activity relationship (QSAR) models offer a computational approach to predict chemical properties.

Purpose of the Study:

• To develop and compare QSAR models for predicting the mean and maximum atmospheric degradation half-life of POPs.
• To evaluate the performance of the linear heuristic method (HM) and gene expression programming (GEP) for this prediction task.
• To identify robust computational tools for assessing the environmental persistence of organic compounds.

Main Methods:

• Development of QSAR models using molecular descriptors derived solely from chemical structures.
• Application of the linear heuristic method (HM) for descriptor selection and linear model building.
• Utilizing gene expression programming (GEP) for non-linear model development and prediction.
• Validation of models using test sets to assess prediction accuracy (r^2 and RMSE).

Main Results:

• GEP models achieved satisfactory prediction accuracy for both mean (r^2 = 0.80) and maximum (r^2 = 0.81) atmospheric degradation half-lives.
• Root mean square errors for GEP models were 0.448 (mean) and 0.426 (maximum) half-life values.
• The heuristic method (HM) was effective in pre-selecting relevant descriptors and building linear models.

Conclusions:

• Gene expression programming (GEP) demonstrates significant potential as a powerful tool for non-linear QSAR modeling in environmental chemistry.
• The developed QSAR models provide valuable insights into the atmospheric persistence of persistent organic pollutants.
• Computational approaches like GEP can aid in the risk assessment and management of environmental contaminants.