AMBER Archive (2006)

Subject: RE: AMBER: ab initio terminology query

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Wed Nov 01 2006 - 10:44:22 CST

> > In the following example: Data was computed at the
> > "B3LYP/6-31++G(2d,2p)//HF/6-31G(d) level of theory"

> Lower level (HF in this case) is optimization,
> higher level (DFT) is the single point property calculation.

I disagree with this. The ethos of the Amber force fields is that a
consistent model is used for the charge calculation and then the individual
parameters are fitted with this charge set in mind. It is extremely
important when calculating new RESP charges that the same level of theory be
used for the charge calculation as was used in the rest of the force field.

In the case of the FF94, 96, 99XX force fields the level of theory used for
the single point calculations was HF/6-31G(d). The reason for this is that
this combination fortuitously underestimates the charges in gas phase giving
charges that are more consistent with being in solvent. Hence this
systematic error is used obtain solvent phase resp charges from a gas phase
simulation of a residue. The FF03 force field uses a different level of
theory for the charge calculation that includes an implicit solvent model to
obtain solvent phase charges. If you use a higher level of theory for the
ESP calculation you will over polarise your molecule when simulating it in
solvent.

As such one should not change the level of theory used for the single point
part of the calculation without a detailed understanding of why it was
chosen to be that in the first place. With regards to the optimisation part
of the calculation I think one could discuss this for ever. Francois can
probably comment here but I think the generally accepted 'optimum' method is
to do the calculation over several conformations and average the resulting
charges. In all cases the single point ESP should be calculated using the
same level of theory as the rest of the force field was fit to.

So with regards to the above statement, if the authors are specifically
referring to RESP charge derivation for an AMBER force field then they mean
they used B3LYP for the optimisation and then HF for the ESP generation.

However, with regards to QM calculations in general I have no idea what the
general consensus is. Ideally I think people should explicitly state what
they did rather than use a confusing and ambiguous notation.

All the best
Ross

/\
\/
|\oss Walker

| HPC Consultant and Staff Scientist |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu