View source for AndyForesterAmber

=Amber Calculations on 1W7S=

Back to main - https://www.ch.ic.ac.uk/wiki/index.php/AndyForester

Full 1W7S amber input is in "though Leap into gassview" folder on the desktop

Issues with this file are:

* slightly non-0 charge on CSY residue
* no water molecules anywhere

==Energy calculations==

These run locally in 2 seconds and compute the energy fine - simple summation  of terms

==Optimisation==

These run on CX1 using 6000MB memory and 1 processor in roughly 4 hours and 5000 steps

Attempt to run the calculation on 2 processors witht 7000MB memory - time should be halfed but CPU time remains around 4 hours.

Issues with the optimisation:

* Not a smooth start to the energy profile - jump from -2 to 200 kJ/mol! - this smooths out

===Visualising whats changed===

* VMD allows overlay of the 2 structures (start and finish) so we can see whats moved
* It also has a RMS fit to calculate some "fit factor"
* Can also print out a distance matrix using 3rd party script in VMD tcl console. This creates a full matrix/surface.
* I and Mike wrote scripts to pull the diagonal out of the file so we are comparing each residue with itself.
* This plots the classic movement per residue - BASED ON CENTER OF MASS.

==Frequency==

The optimised file can be submitted for a frequency calculation using only Amber. Questions to answer about this are:

* Do H and L ONIOM tags get interpreted for Amber only calculations (giving % frequency in the diff regions)?
* Can this calculation be run in parallel NProc=2?
* Does the Chk file load in Gaussview? How large is it?

The initial frequency calculation ran successfully but created a 1.6Gb text file which made reading it difficult.

Answer to the ONIOM tag question:

* Without ONIOM tags - (opt freq) no split between model/real vibrations
* With ONIOM tags - (opt freq) no split between model/real vibrations
* With ONIOM tags - (freq) split between model/real vibrations

It seems Gaussian does recognize the ONIOM split in pure frequency calculations even if no ONIOM method is being used.