![]() Category A peptides can be easily and reliably docked. Results are grouped into categories reflecting docking accuracy. Peptides were docked against the holo structure. Table 1: Results of peptide docking using Glide. This approaches accuracy of brute-force methods like FlexPepDock (63% for this dataset) at a fraction of the computational cost (24 CPU hours for the Glide protocol). Finally, we applied MM-GBSA scoring to each of the predicted poses, further boosting the accuracy to 58%. In addition, we leveraged the previously reported sensitivity of Glide to input conformations and increased sampling by running multiple independent Glide jobs and pooling the results. Applying the optimal combination of these parameters increased accuracy to 41%. To improve prediction, we modified the Glide algorithm by systematically changing all parameters that control how peptide conformations make it through the Glide funnel. Nonetheless, default application of Glide still only resulted in a successful docking of 24% of the peptides in the benchmark set. Instead, we used a different metric, recently proposed for another docking algorithm (FlexPepDock), which determines the RMSD of the backbone atoms of residues at the protein-peptide interface (the iRMSD), and considers docking a success when one of the top 10 scoring peptides has an iRMSD value below 2.0 Å. Indeed, our standard measure for docking success (heavy atom RMSD of the top ranked pose below 1.5Å) turned out to be unattainable for the vast majority of peptides we tried to dock. In a recent publication, Schrödinger scientists systematically analyzed to what extent Glide can be used for peptides, and then made specific changes to the sampling and scoring algorithms to generate a peptide-specific version of Glide.ĭefault application of Glide for peptides turned out to yield poor results compared to small molecule standards. This raises the question of whether a docking program parameterized to work with small molecules can effectively handle larger peptide molecules. The increased use of peptides and peptidomimetics as drugs has created a need for applying docking tools to these types of molecules as well. The ability of Glide to accurately predict binding modes of small molecules in protein structures is well documented. More information on Glide can be found on the Glide product page. In this article, Senior Application Scientist Thijs Beuming discusses the development of a new Glide docking protocol designed specifically for peptides. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |