Review of QSAR models and software tools for predicting developmental and reproductive toxicity

This report provides a state-of-the-art review of available computational models for developmental and reproductive toxicity, including Quantitative Structure-Activity Relationship (QSARs) and related estimation methods such as decision tree approaches and expert systems. At present, there are relatively few models for developmental and reproductive toxicity endpoints, and those available have limited applicability domains. This situation is partly due to the biological complexity of the endpoint, which covers many incompletely understood mechanisms of action, and partly due to the paucity and heterogeneity of high quality data suitable for model development. In contrast, there is an extensive and growing range of software and literature models for predicting endocrine-related activities, in particular models for oestrogen and androgen activity. There is a considerable need to further develop and characterise in silico models for developmental and reproductive toxicity, and to explore their applicability in a regulatory setting.