What is it about?
Ready biodegradability of chemicals is an important value to asses the environmental toxicity. The paper presents a prediction model developed from a dataset of 728 chemicals. Differently from usual QSAR, the new model mines the dataset to extract the structural fragments that are statistically related to ready biodegradability. The model applies a decision tree using these substructures and some expert rules. The obtained accuracy is not worse than the accuracy of BIOWIN (the model developed by the United States Environmental Protection Agency); moreover this new model includes an easily understandable explanation.
Featured Image
Photo by NOAA on Unsplash
Why is it important?
The classification model produces 5 classes that account for the uncertainty of the prediction of the molecule being RB/notRB. It is tested on 874 new compounds, used as the external test, so obtaining a large evaluation.
Perspectives
The interpretability of the models is of concern for regulatory bodies. This model is an answer.
Prof Giuseppina Carla Gini
Politecnico di Milano
Read the Original
This page is a summary of: A new in silico classification model for ready biodegradability, based on molecular fragments, Chemosphere, August 2014, Elsevier,
DOI: 10.1016/j.chemosphere.2014.02.073.
You can read the full text:
Contributors
The following have contributed to this page
