Ugi Precipitation Predictions
Following a recent discussion about the Ugi Master Table, things got sorted out enough that Rajarshi has built a model to predict which of the top 100 molecules from his docking run against falcipain-2 (V2 receptor) are most likely to precipitate the Ugi product. He get 6 hits, 3 of which are part of the training set and 3 have not yet been attempted yet. The new ones are here:
As we get more information in the table, we can start formulating new hypotheses. One of the factors that I have suspected for a while is that the products that precipitated tended to come from less concentrated solutions. Now we are starting to get results from the same Ugi reactions at different concentrations, where this effect should be more obvious.
For example this compound precipitated from a 1 mmol scale reaction with 4 ml of methanol but remained in solution at a 0.25 mmol scale with 0.25 ml of methanol.
This seems counterintuitive because one would expect increasing the concentration should make it easier to precipitate. But I think those solutions are so concentrated that the reagents compete as role of solvent with methanol.
Rajarshi has not yet taken into account the role of concentration but we have discussed it and can more easily do so as we accumulate more results.