In Silico Approaches to Address Compound Attrition
Peter Gedeck, Christian Krämer, Richard A. Lewis · 2015
There are a multitude of biological reasons for compound attrition that can roughly be divided into toxicity (TOX) and failure of efficacy due to poor pharmacokinetics (PK) and medically insignificant targets. This chapter reviews different classes and target fields of in silico approaches that help reduce attrition. If a large amount of data is available, ligand-based and structure-based methods can be combined. Usually, in silico models are good at predicting new compounds that are similar and bad at predicting compounds that are very different from the training set compounds. The impact of ADME and the change in culture on attrition can be seen in industry surveys of attrition rates across the various phases. Some companies have been able to reduce the attrition rate going from clinical candidate to phase 1 by 15% by bringing ADME into the preclinical lead optimization phase.