Addressing toxicity risk when designing and selecting compounds in early drug discovery
Summary In this article, ‘Addressing toxicity risk when designing and selecting compounds in early drug discovery‘, we discuss the application…
Summary
This article discusses a critical issue that the community needs to address address in order to use the predictive models that we build to the greatest effect.
To understand how useful predictive models are in a practical context, we need to understand ‘prior probability distributions’ (or ‘priors’ for short). This article discusses the importance of ‘priors’ in assessing models in different contexts, including selecting or eliminating compounds, prioritising compounds for further investigation, and combining models for different properties to select compounds with a balance of properties.
In all cases, understanding prior probabilities of adverse events makes it easier to make good decisions in drug discovery, so we can more efficiently find high quality compounds to progress.
Citation details
Matthew Segall and Andrew Chadwick, J. Comput.- Aided Mol. Des., 2010, 24, 957–960.
More MPO resources
Comparing chemical series with probabilistic scoring in StarDrop
This example is taken from a project in which screening of a diverse library resulted in hits from multiple chemistries.…
MPO explorer: automatically building a scoring profile with rule induction
In this example we will use the Profile Builder in StarDrop’s MPO Explorer module to derive a multi-parameter scoring profile, based on a CNS data set.
