Hello lavaflow,

Currently planned, we at MGL would send a top proportion of small molecules for a given protein structure to the Levy Lab. The Levy Lab would then create a "complex structure," one ligand pose and its corresponding protein structure, for each ligand, and prepare/submit a BEDAM job to WCG.

The BEDAM job is a free energy simulation that does not rely on grids but the atoms of the complex structure. Most probably 1 BEDAM job, like an AutoDock job, will be split into several work units. Once the BEDAM jobs are complete for a group of top poses for a given protein, the poses would be re-ranked according to the calculated free energies.

For a quick example, imagine that we give them the top 100 ligands for a given protein, and we only had enough money to test 10 compounds. Suppose that we somehow knew that there were only 10 hits (compounds able to affect the protein's activity and may be worthy of further investigation), and all of them were ranked in the top 100 ; however, say the compounds were ranked 10th, 20th, 30th, ..., 100th. If we tested the top 10 ligands at this point, we would only find 1 hit. From what we observed in the SAMPL Challenge, BEDAM has the capability to calculate free energies so that the new rankings would yield at least 5 ligands, instead of 1 ligand, within the top 10. In other words, BEDAM has been observed to enrich Vina dockings by a factor of 5 (I believe it's actually 4.8 reported in our publication).

Hopefully, I've touched on the mechanics or workflow that you wanted me to discuss. Soon, the Levy Lab will have their own website and section of the forum, but until then let me know if there are any more questions.

Thanks,

Dan