November 15, 2002

Structure Calculation with CNS


**********************************************************************************
NOTE FOR INSTRUCTORS (Georgia State and Georgia Tech):

You will need to copy my directory cns under /21/users/pascale/bcmb8190_00.

**********************************************************************************
 

GOALS:

1) Introduction to structure calculation with CNS

2) Introduction to the visualization package InsightII (or rasmol)
 
 

BACKGROUND ON CNS:

CNS : Crystallography & NMR System:

* A new software suite for macromolecular structure determination mainly developed by Dr. Axel Brunger at Yale University.  It replaces the X-plor software developed earlier by Dr. Brunger.

* General system for X-Ray and solution NMR structure determination of proteins, carbohydrates, nucleic acids and their complexes.  Water, ligands,and ions can be included.

* CNS is freely accessible for academic institution in its entirety, including the source code.

* The source code is written in FORTRAN77 for UNIX-based operating systems.

* Most operations are defined throuh tasks and modules written in the "CNS structure determination language"

* Library modules provide for example NMR random coil chemical shifts, molecular parameters and topology database, as well as multi-dimensional probability distributions for preferred rotamers in proteins and nucleic acids.

* Development of new algorithms can be easily done in the CNS language.

* The new HTML interface provides a friendly user interface and allows new-comers to quickly perform simple standard tasks. (see Figure 1)

* It is available for most computing platforms (including SUN and Cray), however it is only supported for a few (SGI, HP, DEC, PC).

* Today we will use the HTML interface for NMR structure calculation.
 

RUNNING CNS:
 

1) Setup environment variables:

at UGA, type:

%setup cns

2) There are two ways to proceed:

a) For interactive CNSsolve session:

%cns_solve :  (use "stop" to get out)

b) For non-interactive CNSsolve session:

%cns_solve <task-file > output-file
 

3) The user first need to modify distributed task files written in CNS language to create their personal task files and then run them. (see figure 3a) There are two ways to proceed:

a) Using your favorite text editor.

b) Using the interactive online HTML interface.

%cns_web

The world-wide web interface provides forms that can be simply filed out and easily converted to personal task files.  Similarly, personal task files can be converted to world-wide-web forms.(see figure 3b)
 

VARIOUS TASKS OF NMR STRUCTURE CALCULATION:

For example, we will use the 24 amino-acid N-peptide

1) Generate structure file for molecular system using sequence information only.

* open the web interface

%cd
%cd bcmb8190/cns
%cns_web

* First lets modify the distributed file generate_seq.inp:

- click on "Input Files", then "General" and "Edit+Help" for the file generate_seq.inp.
- simply enter Npept.seq for the protein sequence file, N122 for segid, N.mtf for the output structure file also click on hydrogen flag, and unclick the bfactor and occupancy flags. Save as generate_seq.inp in your directory bcmb8190/cns.

You will need to create the file Npept.seq (in your directory bcmb8190/cns) which contains the sequence as followed (no carriage return):

GLY SER MET ASP ALA GLN THR ARG ARG ARG GLU ARG ARG ALA GLU
LYS GLN ALA GLN TRP LYS ALA ALA ASN

* Run non-interactively cns_solve

%cns_solve < generate_seq.inp > generate_seq.out

- using the N sequence, a protein topology file (defines structure of amino acids) and a protein parameter file (defines standard bonds and angles) this command will create two files:

generate_N.out: the output file, it is useful for debugging and for other info.
N.mtf:  the molecular topology file: contains the definition of all residues, and atoms (name, chemical type, charge, and mass) as well as the geometries (bond, angles, etc.). You can look at this file in vi or jot, but this file can not be displayed by InsightII or rasmol.

2) Generate an extended file with ideal geometry.

* First lets modify the distributed file generate_extended.inp:

- click on "Input Files", then "NMR Structure calc" and "Edit+Help" for the file generate_extended.inp.

- simply enter N.mtf for the structure file and N_extended.pdb for the output coordinates file. Save as generate_extended.inp in your directory bcmb8190/cns.

* Run non-interactively cns_solve:

%cns_solve <generate_extended.inp > generate_extended.out

The output coordinate file N_extended.pdb can be displayed. Lets view it using InsightII (or rasmol). To run InsigthII:

%setup msi
%InsightII
In InsightII, go to top menu bar "Molecule" then reach "Get" submenu, there enter the following:
For Get File Type: Click on pdb
For MoleFileName: N_extended.pdb
Execute

You can see that you have created an extended structure of the N peptide !

3) Structure calculation using distance geometry and simulated annealing and experimental constraints:

* First lets look at the file dgsa_N.inp, that I have created for you:

- click on "Input Files", then "enter the name of your CNSsolve input file" into the appropriate box:
/21/users/pascale/bcmb8190_00/cns/dgsa_N.inp
 

* Run non-interactively cns_solve:

%cns_solve <dgsa_N.inp > dgsa_N.out  (YOU DON'T NEED TO DO THAT)

Multiple output coordinate files dg_*.pdb and dga_*.pdb (accepted) are created. Lets view one that I have previously calculated (N_struc.pdb) of them using InsightII (or rasmol).
.

_____________________
This week there is no problem set.