Method for superposition of quaternary structures with multiple chains
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.



QSalign is a method for full QS superposition, which leverages the frequent availability of QS information for homologous proteins. It is a two-step procedure to superimpose full QSs. A first alignment and superposition is carried out to establish correspondances between equivalent chains across the query & target structures. Target structures are re-written with a chain order matching that of the query QS and a second round of structural alignment is performed to yield the final TM-score.


Linux or Unix system is required


QSaling uses the program KPAX version 3 to run structure alignments. KPAX executable is supplied with this package.


Step 1.

Download the archives and “use_QSalign”

Step 2.

unzip the files using

tar -xvfz XXXX.tar.gz

Step 3.

Make sure the folder “use_QSalign” contains the following files/folders:


iii) iv) target_file.txt (sample file, to be replaced as needed) v) README.txt vi) folder 'bin' which contains 'kpax' executables vii) folder 'pdb' which contains query and target pdb files from PDB viii) folder 'pdb_pisa' which contains target pdb files from PISA (this is optional, the program will run if this folder is empty) ix) folder 'pdb_split' which contains the pdb files with chains splitted Note: structures in 'pdb_split' folder are used to find potential targets before running a costly QS superposition. A single chain from the query structure is superposed against a single of each potential target. Potential targets are then selected based on this single chain superpostion if the TM-score is above a certain threshold.

Step 4.

Open “” and change the paths of the following folders according to where its is downloaded :

i)   $BU_PDB         (all files from PDB)
ii)  $BU_PISA        (all files PISA, if any)
iii) $BU_SPLIT       (files with chains splitted)

Redundancy in chain names can be found at times in a file provided by PDB, make sure to make the chain names uniq.
Assemblies with renamed (uniq) chain names can be obtained from

All assemblies from PDB, PISA and splitted files used in 3Dcomplex can be obtained from


iv) $T_CUT  – This is the TM-score cut-off that shows how good is the alignment and hence how similar are two structures. A value in the range (0-1) can be specified; 0 – no match , 1 – perfect match (identical structures).  Default value given is 0.65, and we recommend using values above 0.5.

After making the changes save the file “”

Step 5. [OPTIONAL]

By default the folder “BU_PISA” contains only the top solution from PISA, if one wants to use few more solutions from PISA, then run PISA from CCP4 Program suite  ( and copy the assemblies as XXXX_p1.pdb (already present), XXXX_p2.pdb etc. into the folder BU_PISA.

Step 6.

Run  "" with the full path of the query, followed by the number of subunits of the query structure, followed by the target file (containg 4 letter target code and corresponding number of subunits).
    (a sample target file is provided)
    perl /home/XXXX.pdb 4 target_file.txt

Run time depends upon the size of the query, T_CUT, and number of targets, and can take anywhere from few seconds to several hours.

Step 7.

Output generated in the folder “QSalign_results” as “XXXX.QSalign_out”

Gives the list of structures found similar to the query (if any), along with respective TM-Score and few more attributes as generated from KPAX.

A log folder “QSalign_LOG” is also generated containing intermediate superposition results, i.e., before chain re-ordering.
  1. TRY THE EXAMPLE RUN (all required files are supplied):

perl ./pdb/5rub_1.pdb 2 target_file.txt

If you have any problem or suggestion, please contact: OR