IMP Tutorial
for IMP version 2.4.0
|
This example demonstrates the integrative docking protocol by application to the PCSK9 antigen–J16 Fab antibody complex. It uses a command line tool called idock
. For full help on this tool, run from a command line:
First, obtain the input files used in this example and put them in the current directory, by typing:
(On a Windows machine, use 'copy' rather than 'cp'.) Here, <imp_example_path> is the directory containing the IMP example files. The full path to the files can be determined by running in a Python interpreter 'import IMP.saxs; print IMP.saxs.get_example_path('idock')
'.
The input files include the Protein Data Bank (PDB) structures of the isolated J16 Fab antibody and PCSK9 antigen, antibody_cut.pdb
and 2p4e.pdb
, respectively. The protocol can also include one or more additional sources of experimental information. In this case, a set of cross links between the antibody and antigen are provided in the file cxms.txt
. The cross links file has a simple format; each line lists the residue name and chain ID of the antibody and the antigen, followed by the minimum and maximum CA-CA distance in angstroms.
All steps of the integrative docking procedure are carried out by simply running:
The application makes use of the PatchDock program for docking, which must be obtained separately from the PatchDock website. It must be told explicitly where to find PatchDock (either by giving the –patch_dock
option to idock, or by setting the PATCH_DOCK_HOME
environment variable).
The –precision
argument above controls how many candidate dockings are sampled by PatchDock, and can range from 1 (normal) to 3 (high precision). Higher values result in longer run times. The –complex_type
argument is passed to PatchDock, and identifies the type of complex (AA means antibody-antigen in this case). It alters how the docking is performed (for more information, see the PatchDock help).
Once the docking procedure has finished, the primary output file generated is results_cxms_soap.txt
, the first few lines of which look similar to:
# | Score | filt| ZScore |cxms | Zscore |soap | Zscore | Transformation 1 | -1.571 | + | -0.787 | 0.013 | -0.707 | 1435.624 | -0.864 | -1.869 0.676 -2.196 -7.105 -2.340 25.311 2 | -1.245 | + | -0.624 | 0.013 | -0.707 | 1967.947 | -0.538 | -2.811 0.568 -1.826 -3.739 -0.751 31.180
Each line corresponds to one model; the models are ranked by total score, best first. The scores and z-score pairs for each scoring method (just cxms and SOAP in this case) are also shown (only docking solutions that were not filtered out by any of the data sources – i.e. they scored well against every source – are included in this file). The last column is a transformation (3 rotation angles and a translation vector) that transforms the antibody relative to the antigen (the antigen is not transformed).