Strukturrechnung mit CYANA

Latest revision as of 16:18, 16 December 2019

Startup

The structure calculation will be performed on a Linux server that can be accessed by a remote connection to our computer cluster. For the remote connection from a Windows computer, we use the program MobaXterm. This program provides a shell environment in Windows that allows us the access to the remote computer. (If you use a Mac or Linux laptop, please ask us.) To establish the connection to the server please follow the following steps:

• Copy the ZIP file MobaXterm_v7.4.zip to you personal Windows computer and unpack it.
• The folder contains two files, MobaXterm_Personal_7.4.exe (the program executable) and MobaXterm.ini an additional file containing all passwords and data necessary to automaticall access the server
• Start the program by klicking the execuatble MobaXterm_Personal_7.4.exe.
• The program starts with a default shell console (black area). Additionally you see on the left side a field named "Saved session". There, please click on the preconfigured connection 141.2.222.9 to connect. A new tab should open showing the prompt "honshu>". This is the name of our server.

The program shows the list of the files in the active directory on the left side. This enables you to open files in Windows by clicking them and in addition this allows you to copy files between Windows and the active directory on the server by drag and drop.

All students use the same account. As an first step, please create a directory for your group to work in. For that the following steps are necessary:

• Use the command 'mkdir' command to create a directory. Please choose the last name of one of the group members. Use only the standard letters A-Z and a-z. Every group name has to be unique. E.g. if your name is Müller, use 'Mueller' as the groupname, type

mkdir Mueller

• Now change the directory using the command 'cd'. For our example:

cd Mueller

In your group directory, unpack the input data for the CYANA structure calculations:

tar zxf /home/guest/cyanadata.tgz

(It is also possible to download the tgz-archive file cyanadata.tgz with the input data for the CYANA structure calculations.)

Three subdirectories will be created (see README.txt for a description of input and output files):

• manual-manual: structure calculation with manually assigned chemical shifts and manually assigned NOESY peak lists
• manual-noeassign: structure calculation with manually assigned chemical shifts and automatically assigned NOESY peak lists
• flya-noeassign: structure calculation with automatically assigned chemical shifts and automatically assigned NOESY peak lists

Enter the corresponding subdirectory. For example

cd manual-manual

and copy your own data to this directory. For example, for the 'manual-manual' calculation, copy your chemical shift list (sso7d.prot) and your assigned peak list from the 15N-resolved NOESY spectrum (N15NOESY.peaks).

The structure calculation with automated NOESY assignment ('manual-noeassign') can use the same peak list, but CYANA will not use the assignments, only the peak positions and volumes.

CYANA structure calculation with manually assigned data

Start CYANA, and execute the CYANA script CALC.cya (in 'manual-manual'):

honshu> cyana
___________________________________________________________________

CYANA 3.98.12 (linux64-intel)
 
Copyright (c) 2002-18 Peter Guentert. All rights reserved.
___________________________________________________________________

    Demo license valid for specific sequences until 2020-12-31

    Library file "/home/guest/programs/cyana-3.98.12/lib/cyana.lib" read, 41 residue types.
    Sequence file "sso7d.seq" read, 65 residues.
cyana> CALC

In this case, the structure calculation is performed in the terminal window and using a single processor.

For the subsequent structure calculations, it is more efficient and convenient to run them in batch mode on 50 processors in parallel with the command

cyana -n 50 CALC

You can monitor the status of your jobs with the command

qa

(Note that the jobs from all groups are run under the user name 'guest'. Check the directory name in the last column of the output to identify your job.)

Analyze the results of the structure calculation

The results of the structure calculation are the structural statistics in the overview file, sso7d.ovw ('manual-manual') or final.ovw ('manual-noeassign' and 'flya-noeassign'), and the structure itself, which is represented by a bundle of 10 conformers whose coordinates are stored in the PDB file, sso7d.pdb ('manual-manual') or final.pdb ('manual-noeassign' and 'flya-noeassign').

The overview file has at least three parts. The file starts with a table of the target function values and restraint violation statistics. For example:

Structural statistics:

str   target     upper limits    van der Waals   torsion angles
    function   #    rms   max   #    sum   max   #    rms   max
  1     1.69   2 0.0076  0.36   4    5.6  0.34   0 0.3302  3.23
  2     1.74   2 0.0077  0.36   5    5.9  0.34   0 0.3272  3.30
  3     1.75   1 0.0075  0.36   5    5.7  0.34   0 0.3695  3.45
  4     1.87   1 0.0075  0.37   7    6.3  0.34   0 0.3159  2.69
  5     1.95   1 0.0075  0.37   5    6.7  0.37   0 0.3185  3.00
  6     2.12   2 0.0084  0.36   6    6.6  0.34   0 0.3745  3.56
  7     2.19   2 0.0100  0.50   7    6.8  0.34   0 0.3257  3.29
  8     2.35   2 0.0096  0.36   8    7.0  0.34   0 0.3748  3.50
  9     2.40   2 0.0088  0.36   5    8.4  0.35   0 0.4152  3.44
 10     2.49   2 0.0090  0.36   9    7.6  0.33   0 0.3494  3.20

Ave     2.06   2 0.0084  0.38   6    6.7  0.34   0 0.3501  3.26
+/-     0.28   0 0.0009  0.04   2    0.8  0.01   0 0.0308  0.25
Min     1.69   1 0.0075  0.36   4    5.6  0.33   0 0.3159  2.69
Max     2.49   2 0.0100  0.50   9    8.4  0.37   0 0.4152  3.56
Cut                      0.20             0.20             5.00

This table has one row for each structure, containing

• the rank of the structure sorted by target function value
• the target function value
• three columns for each type of conformational restraints:
  • the number of restraints that are violated by more than the cutoff value given in the last row (“Cut”)
  • the root-mean-square (RMS) violation calculated over all, violated and fulfilled, restraints of this type
  • the maximal violation

The five bottom lines of the Table give the average value, the standard deviation, the minimum value, and the maximum value of the corresponding quantity over the individual structures, as well as the cutoff value for significant violations.

Restraints that are violated in a significant number of structures by more than the corresponding cutoff value are reported in the second part of the overview file:

Constraints violated in 3 or more structures:
                                               #   mean   max.  1   5   10
Upper QB    LEU   17 - QB    PRO  108   3.69   3   0.10   0.50  ++    *     peak 1009
Upper HB    ILE   85 - H     ASP   86   3.80  10   0.36   0.37  +++*++++++  peak 803
VdW   N     ILE   81 - HD2   PRO   82   2.45  10   0.34   0.37  ++++*+++++
VdW   CG2   ILE   81 - C     ILE   81   2.90   6   0.20   0.21   + + ++* +
2 violated distance restraints.
0 violated angle restraints.

Each line identifies a violated restraint, and gives the number of structures in which the restraint is violated by more than the aforementioned cutoff value (column labeled “#”), its maximal violation (column “max.”), and the structures in which the violations occur (a one-character column for each structure that is analyzed). Structures in which the restraint is violated by more than the cutoff are marked with “+”, or with a “*” for the structure in which the maximal violation occurs. If available, the number of the cross peak from which the restraint originated is given at the end of the line.

At the end of the overview file, root-mean-square deviation (RMSD) values for the atom positions after optimal superposition of the individual conformers onto the mean coordinates are given:

RMSDs for residues 1..65:
Average backbone RMSD to mean   :    0.58 +/- 0.11 A (0.45..0.82 A; 10 structures)
Average heavy atom RMSD to mean :    1.08 +/- 0.11 A (0.93..1.25 A; 10 structures)

The residue range used for the superposition is indicated. RMSD values are computed for the backbone and heavy atoms of the given residues. The average value, the standard deviation, and the minimal and maximal values of the RMSDs between the analyzed structures and their mean coordinates are calculated.

Visualize the structure

Look at the structure with a molecular viewer such as PyMol. The 'ref.pdb' file contains a "good" structure for comparison.

Structure calculation with automated NOESY peak assignment

The 'manual-manual' structure calculation relies entirely on manual NOESY cross peak assignments. Alternatively, the NOESY cross peaks can be assigned automatically. In the 'manual-noeassign' directory you can calculate a structure from partially or completely unassigned input NOESY peak lists. The input data are:

sso7d.seq

amino acid sequence

N15NOESY.peaks

peak list from 3D ¹³C-resolved NOESY spectrum

C13NOESY.peaks

peak list from 3D ¹⁵N-resolved NOESY spectrum

sso7d.prot

¹H, ¹³C, and ¹⁵N chemical shift list

sso7d.aco

dihedral angle restraints

CALC.cya

CYANA macro (script) for the structure calculation

Combined automated NOESY cross peak assignment and structure calculation are performed with the macro file CALC.cya:

peaks       := N15NOESY.peaks,C13NOESY.peaks    # NOESY peak lists in XEASY format
prot        := sso7d.prot                       # names of chemical shift list(s)
restraints  := sso7d.aco                        # additional (non-NOE) restraints
tolerance   := 0.03, 0.03, 0.5                  # shift tolerances: H, H', C/N', C/N
structures  := 100,10                           # number of initial, final structures
steps       := 10000                            # number of torsion angle dynamics steps
randomseed  := 434726                           # random number generator seed
reference   := ref.pdb                          # reference structure (for RMSD comparison)

noeassign peaks="$peaks" prot="$prot" autoaco

First several variables are set: The variable peaks gives the names of the input peak lists, separated by commas. The variable prot gives the name(s) of the input chemical shift list(s). If a single name is given as in the example, it specifies that a single chemical shift list file with this name will be used for all peak lists. Alternatively, it is possible to specify a separate chemical shift list for each peak list as a comma-separated list of file names. The variable restraints specifies the names of input files with additional conformational restraints that will be used together with the upper distance bounds that will be derived from the NOESY peaks. If there are several file names, they must be separated by commas without intervening blanks. The variable tolerance specifies the tolerances for the matching of chemical shifts. It is used for a consistency check of the peaks that have assignments in the input peak lists, and for the automated NOESY cross peak assignment.

Seven cycles of combined automated NOESY assignment and structure calculation are performed, followed by a final structure calculation. In each cycle and in the final structure calculation 100 conformers are calculated using the standard simulated annealing schedule with 10000 torsion angle dynamics steps per conformer. The 10 conformers with the lowest final target function values are analyzed. An overview table of these 10 best conformers is saved in the file final.ovw, and their coordinates are written to the PDB file final.pdb. The corresponding files from the intermediate cycles 1-7 are called cycle1.*, cycle2.*, etc.

An overview table of the complete calculation can be obtained with the command cyanatable -l (at the Unix prompt) during or after the completion of the calculation.

Structure calculation with automated chemical shift assignment and automated NOESY peak assignment

The 'flya-noeassign' directory provides a structure calculation with automatically assigned chemical shifts and automatically assigned NOESY peak lists. The input data are:

sso7d.seq

amino acid sequence

sso7d.aco

torsion angle restraints (from TALOS)

ref.prot

reference chemical shift list (for comparison only)

ref.pdb

reference structure (for comparison only)

N15NOESY.peaks

peak list from 3D ¹⁵N-resolved NOESY spectrum

\*.peaks

peak lists from other spectra

CALCFLYA.cya

CYANA script to run chemical shift assignment (FLYA)

CALC.cya

CYANA script to run NOE assignment & structure calculation

Automated chemical shift assignment uses, in addition to NOESY, through-bond spectra for backbone and side-chain assignment. The FLYA automated chemical shift assignment calculation is started with the command

cyana -n 20 CALCFLYA

The automatically determined chemical shift assignments are given in the 'flya.prot' output file. The contents of the other 'flya.*' output files are described here.

Using the 'flya.prot' chemical shift list, a structure calculation with automated NOESY assignment can be started, as above, with the command

cyana -n 50 CALC