This discussion is primarily for the 5 teams that submitted their collections of 50 compounds from Enamine: we are asking you to send us a selection of 150 compounds instead of 50. The reason is that over 50% of the compounds in your original selections are predicted to be insoluble at 100 µM (using Simulation-Plus, a predictor recommended by colleagues from Pharma). Computationally selected compounds are typically weak and need to be tested at high concentration (50 µM or higher) were solubility can become an issue. Poorly soluble compounds are likely to form aggregates leading to false positive signals (the assay say compounds are active while they are not). On the other hand, concentrations can be much lower than thought because most of the material has precipitated (the assay says compounds are inactive while they are). Either way, the end result is that the experimental testing is corrupt and the evaluation of the computational method inaccurate. Unrelated but also a problem. A dozen of selected compounds include reactive groups (especially over a handful with an acrylamide handle): we use Surface Plasmon Resonance (SPR) where the protein target is immobilized on a chip and where each compound is let to associate and then dissociate, one after the other. If compound X binds covalently, it will not dissociate from the binding pocket and compounds tested after compound X will not be able to bind since the pocket is occluded. @LucaChiesa will follow-up directly with details with each of you.