Many thanks for this.
I tried the two test.pdb files. They both look fine.
I modified the cluster.in as follows.
trajin prod_0-10ns.mdcrd netcdf
image origin center familiar
cluster out test-out representative pdb \
average pdb averagelinkage clusters 5.0 mass rms :1-123 .CA
The results are pretty awful.
I'm inclined to believe that periodic box info has been lost.
Are there any fixes for this?
Once again, many thanks
George
On Sep 12, 2011, at 9:13 PM, Thomas Cheatham wrote:
>
>> I've been trying to cluster a 10ns trajectory using
>>
>> ptraj solvated.prmtop cluster.in
>>
>> where cluster.in is:
>>
>> trajin prod_0-10ns.mdcrd
>> cluster out test-out representative pdb \
>> average pdb averagelinkage clusters 5.0 mass rms :1-123 .CA
>>
>> Clusters are produced as expected, however, when I try to visualise the
>> clusters (e.g. test-out.c1) using the original solvated.prmtop file
>> produced by tleap, bond lengths, etc. are mangled.
>
> The cluster routine does not "alter" the coordinates per se so there must
> be some mismatch,
>
> try
>
> trajin prod_0-10ns.mdcrd 1 2 1
> trajout test pdb
>
> and look at these. If PDB number 2 is messed up, then it is a prmtop
> mismatch. Alternatively, perhaps molecules have "shifted" due to periodic
> imaging and you need to image prior? [Note also that the average
> structure may of course have funny bonds since motional averaging will
> lead to not representative structures.]
>
> For the "structures" from the clustering, if you looking at a trajectory
> (not the pdb's), box information could be absent.
>
> Note also in that clustering command the space in the cluster mask may
> not be tolerated.
>
> cluster out test-out representative pdb \
> average pdb averagelinkage clusters 5.0 mass rms :1-123.CA
> ^
>
> --tec3
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Sep 12 2011 - 13:30:02 PDT
