John Irwin's blog

A few subsets had duplicates - now fixed

If you downloaded any of the following subsets in the last two weeks, you may have downloaded two copies of each molecule due to a mistake that we fixed today.

The affected subsets are: bioblocks, chembridge BB, princeton BB, florida, otava, otava BB, otava on demand, timtec, ibs, asischem.

The dates are: around Apr 25 - May 6

The impact is: you might have two copies of each molecule.

The remedy is: download a fresh copy.

The worst that can happen is: you see the same molecule twice in your docking calculations.

The severity of this problem is: low

Recent catalog updates

Over a dozen vendor catalogs have been recently updated in ZINC, maintaining our committment to deliver timely information. We update catalogs on a continuous basis, and we try to update every vendor catalog at least once a year, sometimes quarterly.

What does "updated" mean?
1. Compounds that are no longer in the vendor's catalog are marked "depleted"
2. All new eligible compounds are loaded in ZINC.
3. The vendor catalog is exported and publicly accessible in the "by vendor" pages.

"Everything" is ready - final update

Finally, the "everything" subset of ZINC is ready. As of May 7th at noon, it is complete.

Export started on April 21st, so even though ZINC is updated continuously, it is a pretty accurate reflection of the current contents. There are 21,603,031 molecules in the everything subset, which includes all public molecules in ZINC you can buy as well as ones you cannot (e.g. PubChem). Like all ZINC subsets, the "everything" subset is available for download in mol2 and SDF format. The "everything" subset is subset 10, and is available here:

More spring housecleaning

In our ongoing attempt to raise the quality of molecules in ZINC we found 413,986 molecules that were broken in some way. Whereas this may sound like a lot, most of them will have been invisible to you. However, for complete transparency, we are making the file of all these deleted molecules available.

We strongly urge you NOT to use these molecules, because, as far as we can tell, there is something wrong with them.

ZINC (and back again after 18hrs off-line

We had a disk crash over the weekend here. Yes, even RAID-6 disks can crash. Everything appears to be back to normal, but let us know if you notice anything. Started around 6pm Sunday night San Francisco time and was finally resolved within a few minutes of noon Monday morning.

-- John

Spring Cleaning 2009 - deleted 11,471 broken molecules

When we find broken molecules in ZINC we generally try to fix them. However, from time to time we remove molecules that are broken in ways that are not immediately obvious how to fix. A common source of these problems is molecules that were incorrect in the source database, although our scripts are certainly at fault some of the time.

upgrade today - ZINC will be down for a couple of hours

ZINC will be down briefly later today for upgrades. Target window is currently 2 to 5 pm San Francisco summer time, UTC+7, thus beginning 21:00 UTC Thu Apr 16. Total downtime should be under 3 hours. Some services may continue to work.

Two vendor updates for Tue Apr 7

The ZINC entries for Analyticon Discovery (acdisc) and Shanghai Sinofluoro Scientific (sss) were both updated today with the latest information received from the vendors.

We are updating ZINC every day with the latest information from over 80 vendors. We have about 10 more vendors in the pipeline, so we are aiming to go over 100 vendors soon.


Subset #23 (drug-like-now) available today.

A new subset, "drug like now" that can be sourced from shelf stocks, and thus should be deliverable within 2 weeks from most vendors, appeared today.

The "now" subsets of ZINC (21,22,23) are for investigators who do not want to consider those compounds that are "make on request", which can take up to 10 weeks to deliver.

Syndicate content