DOCK Blaster:FAQ

DOCK Blaster FAQ

What is DOCK Blaster?

DOCK Blaster is a free web-based service for molecular docking and virtual screening. It uses UCSF DOCK 3.5.54 as a docking engine, and ZINC as a dockable database. It runs entirely on our 400 CPU cluster in our Mission Bay laboratories in San Francisco. The graphical user interface comprises an expert system that is designed to help you get useful results. We aim to publish a paper in 2007.

Why do this? Docking programs have been around for ages

We are attempting to serve a constituency of investigators who would like to use docking but find the barriers to setting up their own computational lab prevent them from doing so.

How do I use DOCK Blaster?

You need a structure of your target, and an indication of the binding site for which you want to find a small molecule ligand. The target should be in PDB format, and must conform to certain rules (see Preparing Input. The indicatation of the binding site is ideally a docked ligand in mol2 format, but could also be residues surrounding the binding site in PDB format. Point your browser to blaster.docking.org, enter your data, and click "I agree, DOCK!".

How much does DOCK Blaster cost to use?

DOCK Blaster is free to everyone. It is supported by NIH grant GM71896 (to BKS and JJI).

I'm having trouble using DOCK Blaster

We anticipate several sorts of problems:

• 1. The entire site is broken or seems to be behaving incorrectly.

If so, please write to support at docking.org, describing the problem, so we can fix it.

• 2. You cannot get a job to start, because DOCK Blaster complains about your input during the Scrutinizer step.

We know this can be frustrating. We have tried to make the error messages helpful, descriptive and short. Please attempt to follow them. However, if you are not making progress after you have tried 5 times to correct your input, please email support at docking.org, including your JOB ID number if available, so that we can look at the problem.

• 3. Your job gets queued, but does not get started.

Solution: our cluster may be very busy. Please check back again tomorrow.

• 4. Your job fails or seems to get stuck during "Target Preparation".

Solution:

• 5. Your job fails or seems to get stuck during "Calibration".
• 6. You get to "Calibration Done", but you don't know what to do next.

Solution:

• 7. You cannot start a docking job.

Solution:

• 8. Your queued docking job will not start.

Solution: Our cluster may be very busy. Please check back again tomorrow.

• 9. Your job fails or seems to get stuck during "Docking".

Solution:

• 10. You get to the "Results" stage, but you don't know what to do.

Solution:

• 11. You are having trouble with Jmol/Java, Chimera, PyMol.

Solution:

• 12. You have exhausted your quota of allowed jobs for the day.

Solution: please come back tomorrow and try again, or write us at support at docking.org stating why you think you should have a higher quota.

• 13. You ran your job 14 days ago and it has vanished from the system.

Solution: We purge data that has been dormant for 14 days. Please visit and "touch" your job to have it retained for 14 days. 14.

Can I trust DOCK Blaster?

You are wise to be cautious about trusting the results produced by DOCK Blaster. Molecular docking and virtual screening methods are notoriously succeptible to many possible failures due to the numerous approximations made in the calculations. We recognize this inherent skepticism, and have therefore attempted to provide tools with which you can reassure yourself that DOCK Blaster results really are worth investigating - or not, as the case may be.

Here is the evidence we will gather to investigate the calculation:

• 1. score of the supplier ligand in its supplied pose
• 2. ability to re-dock the supplied ligand
• 3. score & pose of similar ligands, drawn from ZINC
• 4. score & pose of property-matched, dissimilar ligands drawn from ZINC
• 5. score of extrema ligands
• 6. score of actives, if supplied
• 7. score of inactives, if supplied
• 8. appearance of top scoring ligands from the lead-like or fragment-like database.

Now, we review how to evaluate the above results in order to estimate whether the results are worth pursuing, for example, by purchasing compounds.

I think DOCK Blaster has crashed on my job

Normally, DOCK Blaster should recover from a crash, and do two things:

• 1. Inform the site manager that a crash occurred.
• 2. Email the originator of the job, if an email address was supplied.

Sometimes this email may not be sent for up to 3 hours after a crash. It is possible that the email sending itself fails.

If you believe your job has crashed, we suggest one of two actions, depending on the nature of the crash.

• 1. If you believe the job crashed due to a general system failure, and not due to your calculation itself, then we recommend trying to re-run the job before contacting DOCK Blaster support. If you try again, and the job crashes at the same or similar point, then please write to support at docking.org to inform us of this so we can investigate.
• 2. If you believe the job crashed due to causes specific to your calculation, then please write support at docking.org right away, telling us any relevant information you think we should know.

I have privacy concerns about my data

DOCK Blaster is a free public service, run on taxpayer-supported computers. However, we do offer limited support for private calculations as follows. When you submit a job, you may provide a PIN. For the first 7 days, this PIN is necessary to access the data in the project. After 7 days, the PIN is removed, allowing public access to your job. Jobs that have not been "touched" for 14 days are removed from the system. Thus, if you want to keep your data private, we suggest deleting from our system after 6 days.

DOCK Blaster made completely wrong suggestions for my target

DOCK Blaster is a research tool aimed at helping people discover new ligands.

Is there any evidence that DOCK Blaster works?

DOCK Blaster uses the same docking technology (DOCK 3.5.54) and database technology (ZINC) as is used in the Shoichet Lab for all their publications. Furthermore, the DUD paper (JMC 2006). Furthermore, the DOCK Blaster paper (to appear).

Why are you guys doing this, I mean, DOCK Blaster?

We felt that many investigators who would like to use virtual screening were not using it because of several barriers to entry. We hope that DOCK Blaster will fill an niche and give access to this technology to investigators who previously were put off by the requirements to set up a docking laboratory.

Can I get my own private copy of DOCK Blaster?

This is something we have though about. If we get enough interest, we may be able to license DOCK Blaster technology. Please contact us to express your interest.

Does DOCK Blaster work better on some targets than on others?

Yes. Now, which ones?

How long does DOCK Blaster take?

Overall, we aim to get you preliminary results in under an hour. Big database docking typically takes a day or so, depending on many factors. Throughput depends on the load on our servers. Thank you for your patience!

What do DOCK Blaster results mean?

Top scoring hits are computer generate predictions of small molecules that might bind to your protein. Our success rate against actual drug targets is on the order of 5% or less, typically. Sometimes it is zero percent! Typically, we purchase 50 compounds that we like from the top 500 for experimental testing, and typically, we find a few that bind in the micromolar range.

Can I used DOCK Blaster with a different docking program?

Currently, DOCK Blaster only supports DOCK 3.5.54. It has not escaped our notice that the overall system may be appropriate for other docking engines. If this option interests you, please write us, making the case for a particular docking program.

Can I used DOCK Blaster with a different database of small molecules?

Yes, there are four ways to pick database subsets to dock.

• our preferred subsets
• annotated subsets
• uploaded subsets
• user-created subsets

We recommend the use of the "fragment-like" and "mini-lead-like" subsets of ZINC.

When does this (awesome) idea become reality?

DOCK Blaster has been in some sort of pre-alpha testing, in one form or another, since 2003. It entered full alpha testing involving 20+ testers on 4 continents in March 2007. Assuming the alpha tests proceed as expected, we hope to make a beta widely available by the fall of 2007, with a view to releasing DOCK Blaster perhaps early in 2008.

I have a question that isn't listed here...

Please use 'search' to check the documentation, or use google, or try to ask our panel of experts in the blaster-fans@docking.org mailing list. If you still can't get an answer, please write support at docking.org.