We recently turned on new compiler flags that caused DOCK to run faster, but also caused an error on a very small number of molecules, perhaps 1 in 10,000. As a consequence, these molecules got hugely negative VDW scores.
This rare bug has been fixed. If you got a hugely negative score (like -1000000 kcal/mol), from circa Sept 1 - Nov 11, now you know why.
The work around is to run the calculation again.
A number of you are waiting for me to follow up on various issues. Please be assured that I have not forgotten you. I regret taking so long to reply. I will try to catch up over the coming weekend (yeah, right).
We noticed that one important feature of DOCK Blaster was not working during an 18 hour window from Wednesday evening to Thursday afternoon.
The problem was, if you were docking a molecule that was not already in ZINC and had to be built on the fly, the job would hang indefinitely.
This same bug affected the "molecule upload" feature of ZINC.
If you have trouble in this time frame, please simply try your job again.
Dear ZINC users
Chemical Block Building Blocks
AsisChem Building Blocks
We have also released new "by property" subsets
clean drug-like #13
We are continuing to put out more updates as we can.
Dear ZINC users
We have updated the following catalogs recently in ZINC (both adds and depletes)
Princeton Building Blocks
TOSLab Building Blocks
A dozen more updates are in progress and will be reported soon.
We just got the ZINC upload feature working again today. It has been down since the cluster outage on the Nov 1st weekend. Thanks to colleagues in Mattia Falconi in Rome, Italy for bringing this to my attention.
Please tell me if something does not work. It helps me prioritize.
New Fragment-like (#2) and lead-like (#1) subsets of ZINC were released yesterday (Nov 3, 2009). These subsets are updated to include over 200,000 depletions (compound no longer for sale) and over 300,000 new compounds (newly on the market).
We are midway though our fall update. These subsets will be updated again before the end of 2009. Updates to other subsets and catalogs are in progress and will be announced as they become available.
Sometimes people ask us why we think they should use DOCK and not some other docking program. Of course, we openly admit our bias. DOCK is the program we develop and the one we know best. But that is not all. We have put together a short presentation of why we think you should consider using DOCK for your next ligand discovery project.
Our Nature paper has appeared. Please check it out!