run.cns

The file run.cns all parameters for the ARIA, the simulated annealing script, and the analysis are stored in the following compound parameters:

Spectrum

the name of the spectrum currently worked on

Iteration

the iteration currently worked on

Filenames

various filenames (e.g., psf file etc)

Spectra

parameter for spectra (name, mixing time etc)

Data

parameter for data

Iterations

aria assignment parameters

Saprotocol

simulated annealing protocol parameters

Refine

refinement parameters (e.g., h2o refinement)

Relax

relaxation matrix calculation

Toppar

topology and parameter information

Analysis

analysis parameters

The following variables are important for the NOE assignment process:

Qshift

use re-assignment based on chemical shifts. If true, shifts file is read in and used for the assignments. If false, only fully assigned peaks will be used.

Qrelax

switch spin diffusion correction off or on (if true, you need to set spectrometer frequency, rotation correlation time, mixing time)

Qcalib

use automated calibration to re-derive target distances from volumes. If true, calibration is switched on. If false, ARIA does not calibrate the peaks.

Qerrset

re-define error bounds based on target distance or volume.

Qmove

re-``calibrate'' violated restraints by setting them to 0...6 (if true, moving of upper bounds to 6.0 A possible)

Qexclude

remove violated restraints (exclusion of peaks possible/impossible)

qfadjust

not yet implemented

errmod

DIST or VOLUME (at the moment, we use DIST as default)

solvent

not yet implemented

Please have a look at the atomname nomenclature of your peak lists (use the atomname nomenclature switch in run.cns). ARIA writes pdb files in IUPAC format (thus e.g. LEU HB2 and HB3 instead of HB2 and HB3). The termini are H1, H2, H3 and OXT, O. Please have a look at topallhdg5.2.pro in the toppar directory to make sure that you are using the same conventions.

You can use CSI or TALOS restraints with our setup, but also be aware of the empirical content of these restraints. You can switch them on or off in run.cns.

You may want to play around with the maximum number of assignment possibilities maxn, the violation tolerance, the ambiguous cutoff and the violation ratio (see [7] for a discussion of these values).

Usually, we calculate 20 structures, take the 10 best structures (regarding total energy) as template structures for the next iteration and the 7 best structures for the statistics upon which we accept or reject peaks.

For setting up aria on several machines or with a queuing system, please specify all the cns executables for all the machines you are using (e.g. a SGI and a linux executable) and use something like csh or rsh or ssh or dqs to distribute them to the other machines. The cns jobs are distributed using some .job files (see temporary directory) which are started with the specified queue command. The Python jobs will always run on the machine on which the main job is running.

You may want to change the number of steps of your simulated annealing protocol. To improve convergence, please multiply the number of steps by 2, 4 or even 8. The default values are pretty good, but you may change them as you like. Water refinement is switched on by default. It should be done before submission of your structures to the pdb.

You can specify procheck, whatcheck and prosa executables in the run.cns files (it will produce some nice statistics in ascii and LaTeX format).

ARIA sets some default values automatically. Thus, a first try would be to calculate some structures without touching the run.cns file at all. More experienced users may want to edit the assignment parameters and the simulated annealing parameters at will. You can also change the CNS protocols, e.g. refine.inp.