ADL: error message in autodock4

Garrett M. Morris garrett at scripps.edu
Wed Mar 26 18:49:20 PDT 2008


Dear Borries,

On 26 Mar, 2008, at 5:53 am, Borries Demeler wrote:
> Hi,
>
> I didn't search your archives to see if this was mentioned before,
> but I didn't see it listed in your faqs...
>
> When screening a large library autodock4 fails to converge in
> about 0.5% of the compounds listed. In those cases, I see the
> message:
>
> NOTE: All energies are equal in population; re-initializing....
>
> endlessly repeated in the log file, which grows very large. I am not
> sure if this is a bug in the initialization, but in my opinion the
> program should exit in such a case more gracefully. I never see this
> string in any of successful fits, which would indicate to me that
> the program has hit en error condition from which it can't escape.
> So my suggestion would be to better deal with the initialization, or
> change the output message to better reflect the true error condition,
> whatever it is.
>
> I should also mention that some of the compounds miraculously do
> converge if I just restart the program. I don't know how the program
> works, but my interpretation of this symptom would be that you use
> time to initialize the random seed of your GA, and then when I restart
> it I get a better initialization. However, when the program gets stuck
> in this loop, even though it says that it reinitializes, it never  
> actually
> does. Does that seem plausible?
>
> Thanks, -Borries
> ________________________________________________
> --- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list  
> ---

This has been fixed in the current unreleased version of AutoDock; in  
the meantime, the workaround for these problem molecules is to use  
"unbound 0.0" instead of "compute_unbound_extended" in these  
particular ligand's DPFs.  (Note: you might also try adding  
"include_1_4_interactions" to the start of the DPFs, this also might  
resolve these cases.)

Please accept my apologies for any inconvenience this may have caused.

Garrett

___
Dr Garrett M. Morris, M.A. (Oxon.), D.Phil. (Oxon.)

The Scripps Research Institute,       tel: (858) 784-2292
Dept. Molecular Biology,  MB-5,       fax: (858) 784-2860
10550  North Torrey Pines Road,       email: garrett at scripps.edu
La Jolla,  CA 92037-1000,  USA.       mgl.scripps.edu/people/gmm






More information about the autodock mailing list