Predictive Models

Asteris Predictive Models

In silico predictive models provide estimates of a compound’s properties based only on its structure and can be used to guide the selection and design of compounds before synthesis or testing. Asteris provides a suite of predictive ADME and physicochemical models that can be used to explore a wide range of compound options prior to selecting a synthetic strategy.

Using Asteris you can calculate a range of simple “core properties”, such as logP, molecular weight, polar surface area and numbers of hydrogen bond donors and acceptors, and predict a number of key ADME properties, including solubility, hERG inhibition and CNS penetration, using rigorously validated models from the StarDrop platform.

Predictive models capture information about the relationship between the structures of compounds and their properties. Asteris’ models help to visualise these structure-activity relationships by providing a Glowing Molecule™ representation for each prediction, indicating the parts of the molecule having the greatest influence on the predicted property. This helps to guide the redesign of compounds by targeting the regions and modifications that are most likely to improve a property.

Note: For some properties an increase may be desirable (e.g. solubility) and for others it may be undesirable (e.g. hERG pIC50) and vice versa.