Structure-based binding pocket detection and druggability assessment

Abstract

Target-focused (often described as ?rational?) drug discovery starts with the identifcation and validation of one or more target biomolecules whose modulation will interfere with disease progression, or at least alleviate disease symptoms. In addition to verifying the association between such biomolecules and the disease, target validation often involves the assessment of some relevant properties, druggability among them.Druggability is often defned as the ability of a target to bind a small, druglike molecule with high affnity and specifcity. If a small molecule has already been reported to bind to a target in a specifc manner, the target is traditionally regarded as druggable (recently, thus, it has been proposed that different targets might present different degrees of druggability, with some of them being ?easy to drug? and some of them being ?diffcult to drug?). Druggability has also been implied based on evolutionary relationships (using a sort of transitivity or ?guilt by association? principle), or by predicting druggable pockets from different levels of structural information, without comparison with any template. The latter approximations, which will invariably depart from the detection of binding pockets and later assign a score to differentiate druggable from non-druggable proteins, are the scope of the present chapter.Importantly, current methods to assess target druggability from structural data might be challenged by the occurrence of hidden or cryptic binding sites, which could more appropriately be assessed by studying the target dynamically, rather than from static representations of the biomolecule of interest.