Re: Biotin force field

From: Kenno Vanommeslaeghe (kvanomme_at_rx.umaryland.edu)
Date: Thu Oct 03 2013 - 13:28:51 CDT

Given the non-response on the past posts you're citing, I doubt you'll get
a response this time; 1997 is 15 years ago. More importantly, you may not
want to use those parameters. I have multiple issues with Izrailev et
al.'s parameter assignment methodology:
* "A charge distribution for biotin was obtained by means of the program
GAUSSIAN-94 (...)": the default charges produced by Gaussian are Mulliken
charges, which are about the worst possible choice for putting directly
into a force field. See
http://mackerell.umaryland.edu/~kenno/cgenff/faq.html#esp
* "... using the coordinates of heavy atoms taken from the crystal
structure (chain B from entry IAVD in the Protein Data Bank) ...": the
coordinates of small molecules in protein X-ray structures are notoriously
unreliable, except if the resolution is very high, which is not the case
(2.7A). Doing a single-point QM calculation on such coordinates is likely
to result in inaccurate properties.
* "The equilibrium bond length, angles, torsional angles, and force
constants for biotin were derived from biotin coordinates and force-field
parameters of molecules with similar chemical structure available in
CHARMM and X-PLOR." The biotin coordinates from a single low-resolution
X-ray structure are all but irrelevant for the purpose of force field
development. Also, in 1997, the CHARMM force field didn't contain any
functional groups that were even remotely similar to the ones in biotin,
so the analogy would be too bad to justify doing this without parameter
optimization.

Don't get me wrong; I'm not saying Izrailev et al. did bad work. They made
do with the force fields, know-how, data and methods at their disposal *IN
1997*. What I'm saying is that their methodology is woefully obsolete by
2013 standards. The CHARMM General Force Field (CGenFF) is now available,
which has much, much better support for biotin (albeit still not perfect).
I'm almost certain a simple initial guess produced by the CGenFF program
will do better than Izrailev et al.'s model.
http://mackerell.umaryland.edu/~kenno/cgenff/program.html
And there's always the possibility of improving the parameters...

On 10/03/2013 11:44 AM, Zack Scholl wrote:
> Dear NAMD community,
>
> Can someone please pass along the CHARMM parameters and topology for
> biotin that were published in Izrailev et al., 1997 (Biophys. J.)? The ftp
> site that maintained them is no longer active.
>
> Note:
> I realize this has been a topic of previous threads
> (http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2008-2009/2675.html,
> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/3338.html,
> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/7801.html) but it
> seems that these parameters are still not available on the net.
>
> Thanks a lot for your time!
> Sincerely,
> Zack
>
>
>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:23:48 CST