The use of in silico molecular modelling to screen potential estrogen mimics as part of medicines and agrochemicals development and product license applications.

Estrogen mimics are a diverse group of synthetic and naturally occurring compounds that can interact with estrogen receptors (ERs) in animals, including humans. These interactions rely on key structural features of 17bestradiol (E2) and if these molecular features are in a similar spatial arrangement on other compounds, they are likely to elicit an agonist (i.e., they are E2 mimics) or antagonist response. The structural diversity of some compounds vis-`a-vis analogies with E2 makes it difficult to reliably predict E2 mimicry on simple structural grounds alone. We propose a new approach methodology: in silico molecular modelling augmented by an in vitro transactivation reporter gene assay to predict E2 mimicry and thus further reduce regulatory reliance on animal studies. Transactivation reporter gene assay dose response curves and in silico molecular modelling were used to obtain EC50-values and docking parameters (DockScores), respectively of thirty E2 mimics to assess the reliability of in silico receptor interaction parameters to predict E2 mimicry. A linear relationship (R2 = 0.75) was found between DockScores and EC50s, suggesting molecular modelling is a good tool for predicting E2 mimicry in a regulatory setting.