If computer memory is really the problem you could split your script
into three individual calculations but I would think that a PCA on 4500
atoms should fit in far less than 1 GB.
What strikes me in your script, however, is that you obviously
attempt to compute quasiharmonics from a mere 30 points. To get all
frequencies you would need at least 3N frames in your trajectory where
N is the number of atoms you choose for the quasiharmonic calculation.
Also, global motions converge very slowly (do a PCA to check) and
entropy computed in that way may be very unreliable (a 1500 residue
system would be quite a challenge). Also, the entropy is dominated by
low frequencies. Reducing the system size will only produce higher
frequencies.
Cheers,
Hannes.
On Wed, 26 Oct 2011 17:53:24 +0800
Neha Gandhi <n.gandhiau.gmail.com> wrote:
> Dear Amber_users,
>
> I agree that the system is too big.
>
> I tried with CA,C and N atoms but no luck! Here is my script. Your
> input is appreciated.
> --------------------------------------------------------
>
> trajin prod36.mdcrd 1 50 10
> trajin prod37.mdcrd 1 50 10
> trajin prod38.mdcrd 1 50 10
> trajin prod39.mdcrd 1 50 10
> trajin prod40.mdcrd 1 50 10
> trajin prod41.mdcrd 1 50 10
> trajin prod42.mdcrd 1 50 10
>
> rms first .CA,C,N
> matrix mwcovar name comp.matrix :1-1471.CA,C,N
> analyze matrix comp.matrix out complex_entropy.out thermo reduce
> matrix mwcovar name rec.matrix :1-894.CA,C,N
> analyze matrix rec.matrix out receptor_entropy.out thermo reduce
> matrix mwcovar name lig.matrix :895-1471.CA,C,N
> analyze matrix lig.matrix out ligand_entropy.out thermo reduce
--
Scanned by iCritical.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Oct 26 2011 - 03:30:02 PDT
