Abstract
Over the past 8 years, we have developed, refined and applied a fragment based discovery approach to a range of protein targets. Here we report computational analyses of various aspects of our fragment library and the results obtained for fragment screening. We reinforce the finding of others that the experimentally observed hit rate for screening fragments can be related to a computationally defined druggability index for the target. In general, the physicochemical properties of the fragment hits display the same profile as the library, as is expected for a truly diverse library which probes the relevant chemical space. An analysis of the fragment hits against various protein classes has shown that the physicochemical properties of the fragments are complementary to the properties of the target binding site. The effectiveness of some fragments appears to be achieved by an appropriate mix of pharmacophore features and enhanced aromaticity, with hydrophobic interactions playing an important role. The analysis emphasizes that it is possible to identify small fragments that are specific for different binding sites. To conclude, we discuss how the results could inform further development and improvement of our fragment library.
| Original language | English |
|---|---|
| Pages (from-to) | 603-620 |
| Number of pages | 18 |
| Journal | JOURNAL OF COMPUTER-AIDED MOLECULAR DESIGN |
| Volume | 23 |
| Issue number | 8 |
| DOIs | |
| Publication status | Published - Aug 2009 |
Keywords
- Fragment screening
- Fragment based drug discovery
- Library design
- Chemical space
- MOLECULE-BINDING-SITES
- PROTEIN-LIGAND COMPLEXES
- DRUG DISCOVERY
- AUTOMATIC IDENTIFICATION
- KINASE INHIBITORS
- NMR-SPECTROSCOPY
- PDBBIND DATABASE
- LEAD GENERATION
- AFFINITY
- DRUGGABILITY