Hello Amber community,
I am trying to calculate the H-H spectral density for amyloid beta by
looking at the correlation function of all H-H distances from 20 ns of NVE
simulation. There are 308 hydrogens in the 42-residue fragment, resulting
in 47278 possible H-H pairs.
I've written a bash loop to insert a "vector" and "timecorr" command for
every such H-H pair, resulting in a ptraj script with essentially 50000
actions in it. The script runs fine, but there are some considerations that
I was hoping you could advise me on.
1) Is there a limit to the number of actions I should specify in one
script? These scripts seem to execute just fine, but perhaps there is some
issue I am missing.
2) cpptraj by default prints out all the actions specified in the input
script. That means that before it runs it has to print close to 300,000
lines (5x per vector & timecorr). Is there a way to make it less verbose?
It takes longer to print out the actions that to actually run them.
3) I use dplr and dplrout commands to print out the extra information, such
as R^-6, but that means the output files have a lot of extraneous
information. Is there an easy way to just print some of the aspects of the
resulting datasets? _TC_DIPOLAR_ is a single dataset, so it's easy to
control, but every timecorr command creates a new TC_##### dataset. Is
there a clean way to subset them all without adding another 50,000 write
actions?
4) Am I doing all of this completely backwards from the get go? I am
following the principles outline here (
http://link.springer.com/article/10.1023%2FA%3A1011241030461) which use
correlation functions, but it seems like the cpptraj functions rotdif can
potentially yield the same information. Do you have any recommendations on
which method I should proceed with?
Thank you,
Eugene Yedvabny
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Oct 30 2014 - 10:00:03 PDT
