Date: Mon, 29 Jul 2002 17:27:52 -0700
From: "David A. Case"
Subject: Re: SANDER documentation for nmropt

On Mon, Jul 29, 2002, a user wrote:
>
> A colleague of mine was trying to run a temperature
> ramp simulation with AMBER 7 and could not get it to run successfully. When
> he showed me his input files, I saw that he had done everything right except
> one thing. He didn't include the
> &rst
> iat(1)=0,
> &end
> lines. After we included these lines in the input file, below his
> &wt...&end lines, the simulation ran successfully.
>
> After reading through the AMBER 7 manual, I had a hard time finding
> somewhere that mentioned the necessity of these lines. Does the manual say
> anywhere that if you specify nmropt=1, that you must also include the &rst
> section, particularly the terminating "iat(0)=1" part, even when there are
> no weight restraints? Have you ever considered adding a warning about this
> under section 5.6.1. General flags describing the calculation?

In Section 5.9, the manual states: "namelist &rst is read repeatedly until a
namelist &rst statement is found with iat(1)=0, or until reaching the end of
the DISANG file." But I agree that this could easily be missed.

Our intention in Amber 7 was to (more or less) force people to use a separate
file (the "DISANG" file) for restraints, rather than putting them into the
mdin file. [See, for example, section 5.5]. Unfortunately, a mistake was made
on p. 111 about this. We fixed this on the web, (see the p.111 entry at the
Amber web site -> Amber 7 manuals -> updates to the manuals) and in the
"second printing" of the manual. But I admit this might be pretty hard to
find, and the above error is understandable.

So, if the restraints are in a separate file, the iat(1)=0 card is not
necessary: end-of-file signals end-of-restraints. If you have no restraints,
it's probably still best to create a zero-length dummy file, and point to
that.

Bottom line is that you are correct: we need to update the instructions (and
examples) to make all of this clearer. I'll post this reponse to the mail
reflector, in hopes that it will help others avoid the same mistake.

..thanks for your input....dac