Very good ideas! I am checking for vaccuum bubbles because I have little
confidence in density actually pickup up problems unless they are VERY
severe, and I already have code that colors a grid in such a way as to
make cavities detectable. I use the code at the moment to squeeze extra
solvent particles into a simulation wherever they might best fit--for
instance, I can tell the code to stay at least 3A from any protein atom,
and insert particles if spots free space least 2A from any solvent residue
(partitioning "protein" and "solvent" pretty much the way it's done in the
prmtop file). Periodicity is taken into account, as is possibility of a
triclinic cell. If the desired number of solvent molecules are not
squeezed in after like 1,000,000 consecutive failures I can order the
program to iteratively reduce the padding for "solvent" particles until it
does get everything in. All of this pertains to crystal space simulations
where you have to worry about just the right amount of solvent to maintain
the proper cell volume.
Dave
On Wed, 3 Oct 2007, Carlos Simmerling wrote:
> you also might want to add checking of HIS residue names,
> density of the system (maybe easier than actual bubbles, which
> would presumably involve looking at volumes/surfaces?),
> check the termini to make sure they are reasonable (charged
> or blocked), check chiralities (very important!!), cis/trans peptide
> bonds.
> carlos
>
> On 10/3/07, David Cerutti <dcerutti.mccammon.ucsd.edu> wrote:
>>
>> Thanks for the clarification, Carlos and Ross.
>>
>> I'm actually working on a "quality control" code that will automate a lot
>> of the things I do to make sure that I haven't made mistakes in my
>> simulation setup or MD input file. It will read and parse a topology,
>> read a set of coordinates, and then tell you if things look all right. It
>> should be easily expandable to include other checks, and the stuff I plan
>> to include should run in about a minute.
>>
>> Currently, the diagnostics I plan are:
>>
>> - Check to see that the water (if water is in the simulation) is normal
>> and alert the user if it's not (SPC/E charges on TIP3P water, for
>> instance).
>> - Check for possible missing disulfide bridges.
>> - Check for vacuum bubbles (very useful if running NVT simulations) and
>> report any such voids in a PDB-format file.
>> - Create a human-readable description of every distinct residue type to
>> help users see the residues from a different perspective and thereby
>> ensure that AMBER understood what they meant if they've introduced
>> something new.
>>
>> In the future, we might want to include diagnostics for:
>>
>> - Minimum distance between two solute images
>> - The net force on each atom due to particular, nearby solute images
>>
>> I'm open to other suggestions as I write this code for myself. I expect
>> that it will save me time and improve my simulations in the long run.
>>
>> Dave
>>
>> -----------------------------------------------------------------------
>> The AMBER Mail Reflector
>> To post, send mail to amber.scripps.edu
>> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>>
>
>
>
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Oct 07 2007 - 06:07:13 PDT
