Optimal selection of learning data for highly accurate QSAR prediction of chemical biodegradability: a machine learning-based approach.

Prior to the manufacture of new chemicals, regulations mandate a thorough review of the chemicals under risk management. This review involves evaluating their effects on the environment and human health. To assess these effects, a review report that conforms to the OECD Test Guidelines must be submitted to the regulatory body. One of the essential components of the report is an assessment of the biodegradability of chemicals in the environment. In addition to conventional methods, quantitative structure-activity relationship (QSAR) models have been developed to predict the properties of chemicals based on their structural features. Although a greater number of chemicals in the learning set may enhance the prediction accuracy, it may also lead to a decrease in accuracy due to the mixing of different structural features and properties of the chemicals. To improve the prediction performance, it is recommended to use only the appropriate data for biodegradability prediction as a training set. In this study, we propose a novel approach for the optimal selection of training set that enables a highly accurate prediction of the biodegradability of chemicals by QSAR. Our findings indicate that the proposed method effectively reduces the root mean squared error and improves the prediction accuracy.

Hazard Evaluation Support System (HESS) for predicting repeated dose toxicity using toxicological categories.

Repeated dose toxicity (RDT) is one of the most important hazard endpoints in the risk assessment of chemicals. However, due to the complexity of the endpoints associated with whole body assessment, it is difficult to build up a mechanistically transparent structure-activity model. The category approach, based on mechanism information, is considered to be an effective approach for data gap filling for RDT by read-across. Therefore, a library of toxicological categories was developed using experimental RDT data for 500 chemicals and mechanistic knowledge of the effects of these chemicals on different organs. As a result, 33 categories were defined for 14 types of toxicity, such as hepatotoxicity, hemolytic anemia, etc. This category library was then incorporated in the Hazard Evaluation Support System (HESS) integrated computational platform to provide mechanistically reasonable predictions of RDT values for untested chemicals. This article describes the establishment of a category library and the associated HESS functions used to facilitate the mechanistically reasonable grouping of chemicals and their subsequent read-across.

Categorization of nitrobenzenes for repeated dose toxicity based on adverse outcome pathways.

Adoption of the data-gap filling method for complex endpoints such as repeated dose toxicity (RDT) and reproductive/developmental toxicity is one of the most important issues affecting international chemical management at present. A categorization method based on adverse outcome pathways (AOPs) has recently been investigated for such complex endpoints. In this paper, we report results of the categorization of nitrobenzenes for RDT based on the AOPs obtained by analysing the detailed RDT test reports for 24 different nitrobenzenes already evaluated. In most RDT testing of nitrobenzenes without hydroxyl groups or acid groups, findings related to haemolytic anaemia and liver effects were observed at low dosages. It was, therefore, possible to assume common AOPs for haemolytic anaemia and liver effects induced by these nitrobenzenes. As a result, a group of nitrobenzenes was defined as a single category for both haemolytic anaemia and liver effects, respectively, based on these AOPs.

Category analysis of the substituted anilines studied in a 28-day repeat-dose toxicity test conducted on rats: correlation between toxicity and chemical structure.

In order to establish methods for estimating the repeat-dose toxicity of chemicals on the basis of their chemical structure, an analysis of a category formed for 14 substituted anilines was conducted. This analysis was based on the results of a 28-day repeat-dose toxicity test conducted on rats in which these 14 chemicals were studied. The intensities of the toxicological effects of the 14 substituted anilines on each target organ at specific dosages were described using the values and histopathological findings of the test. The results clarified the characteristics of the chemical structure that induced specific toxicological effects on specific targets at a particular dosage. Hemolysis was the most frequently observed finding in the test reports in the case of the 14 substituted anilines. Strong linear correlations between the dosage and proportion of decrease in the erythrocyte count were found in the case of chemicals that induced strong hemolytic effects. In particular, for dimethylanilines, strong linear correlations were found between the calculated hemoglobin-binding index and the proportion of decrease in the erythrocyte count at a particular dosage. Thus, the results of our analysis demonstrate that it is possible to correlate the values obtained for substituted anilines from 28-day repeat-dose toxicity tests with their quantitatively determined molecular properties. The intensity of hemolysis and the effects on the liver tended to be low in the case of chemicals with a high water solubility, such as aminophenols and benzene sulfonic acids. However, a similar trend was not observed in the case of the effects of these chemicals on the kidney.

External validation of the biodegradability prediction model CATABOL using data sets of existing and new chemicals under the Japanese Chemical Substances Control Law.

External validation of the biodegradability prediction model CATABOL was conducted using test data of 338 existing chemicals and 1123 new chemicals under the Japanese Chemical Substances Control Law. CATABOL predicts that 1089 chemicals will have a BOD < 60% while 925 (85%) actually have an observed BOD<60%. The percentage of chemicals with an observed BOD value <60% tends to increase as the predicted BOD values decrease. In contrast, 340 chemicals were predicted to have a BOD > or = 60% and 234 (69%) actually had an observed BOD > or = 60%. The prediction of poor biodegradability was more accurate than the predictions of high biodegradability. The features of chemical structures affecting CATABOL predictability were also investigated.