Requires PERL (v.5.8 or higher), and a fortran90 compiler (prefered: ifort) and six additional packages to be pre-installed
- sc from ccp4 (http://www.ccp4.ac.uk/) [executable_name: sc]
- delphi v.6.2 (http://compbio.clemson.edu/delphi) [executable_name: delphi95]
- Rosetta (https://www.rosettacommons.org/software/)
- ProQ2 (https://github.com/bjornwallner/ProQ_scripts/)
- needle from EMBOS (http://emboss.sourceforge.net/download/)
$ git clone https://github.com/bjornwallner/ProQDock
$ cd ProQDock
$ chmod +x compileF
$ ./compileF <fortran90-compiler> (Default: ifort)open ProQDock in any text editor and change the fullpaths for the following fields according to your system:
- sc_path=/software/apps/ccp4/ccp4-6.5.0/ccp4-6.5/bin
- delphi_path=/home/x_sabas/bin
- rosetta_path=/proj/wallner/users/x_bjowa/github/Rosetta/main/source/bin
- rosetta_db=/proj/wallner/users/x_bjowa/github/Rosetta/main/database
- svm_path=/home/x_sabas/proj/svm_light
- ESpath=/home/x_sabas/proj/EDTSurf
- proqpath=/home/x_sabas/proj/BACKUP/ProQ_scripts/bin
1. The coordinate (PDB) file for the model
2. The full-length target sequence
- The model must be a perfect or imperfect subset of the target sequence
- i.e., Any and all sub-sequence present in the model must be present in the tareget
- PDB file MUST contain corrdinates of geometrically fixed Hydrogen atoms
- preferably fixed by REDUCE v.2 or atleast compatible with the REDUCE format (http://kinemage.biochem.duke.edu/downloads/software/reduce/)
The full length (FASTA) target sequence (one letter amino acid code followed by a header starting with >) should be prepared and provided alongside the coordinate (PDB) file for the model.
You can generate the fasta sequence by using:
$ ProQDock/EXEC/pdb2fasta.pl <input.pdb> > input.pdb.fasta$ ProQDock/ProQDock <model.PDB> <fulllength.fasta>where,
-
model.pdb: The input pdb (coordinate file in Brrokheaven format; http://www.ccp4.ac.uk/html/procheck_man/manappb.html) file
-
fulllength.fasta: The fulllength fasta sequence of the target
-
optional arguments: -d (0/1) : delphi_mode:
0: Single_internal_dielectric (epsilon=2) 1: Multidielctric mode (internal dielectric modeled as a Gaussian Function)
If you intend to Run ProQDock for a list of PDB files (models) corresponding to the same target sequence, then, SKIP repeating the profile generation step more than once, by providing a [consistent basename] for the fasta file.
Launch the following bash script which runs on a set of (example) models kept at ./EXAMPLE_CASE_STUDY/
$ ./runex.bashEXAMPLE OUTPUT:
$ cat EXAMPLE_CASE_STUDY/D1OYV-models.status.results Model fnat LRMS iRMS CAPRI_class DockQ ProQDock
D1OYV-a102a 0.000 35.348 15.846 Incorrect 0.02118 0.05908
D1OYV-a104c 0.000 44.703 16.626 Incorrect 0.01432 0.21024
D1OYV-a106c 0.000 44.600 18.136 Incorrect 0.01395 0.03890
D1OYV-a115c 0.000 31.946 11.941 Incorrect 0.02722 0.02730
D1OYV-a61c 0.652 6.820 2.334 Acceptable 0.51761 0.51060
D1OYV-a61d 0.630 6.574 2.450 Acceptable 0.50960 0.46046
D1OYV-a81d 0.587 6.788 2.481 Acceptable 0.48841 0.42648
D1OYV-a41b 0.717 3.619 1.428 Medium 0.69617 0.48751
D1OYV-a62c 0.707 3.991 1.628 Medium 0.66168 0.55795
D1OYV-a82d 0.707 4.347 1.593 Medium 0.65639 0.48664
Example Output for a single PDB file:
$ ./ProQDock ./TESTPDBS/1u07.pdb ./TEMP/1u07.pdb.fastaThe file has a Proper extension (.pdb / .PDB)
=============================================================
Normal Delphi (single internal dielectric, epsilon = 2.0) will be executed
Setting CCP4_SCR to /tmp/tmp.hAw2aARLGb/ccp4_scr!
Cleaning of the temp files in this
directory is not done automatically.
===========================================================
Total number of residues in the Complex: 178
=================================================================================================================================
Feature Range Direction Description Feature_Type Stability* Native_values (DB3)
---------------------------------------------------------------------------------------------------------------------------------
EC: [-1,1] Positive Electrostatic balance at the interface Interface Variable 0.20 (+/-0.38)
Sc: [-1,1] Positive Geometric / Steric fit at the interface Interface Stable 0.68 (+/-0.17)
rGb: [-1,1] Positive Distribution of hydrophobicity with respect to burial All_Atom Stable 0.06 (+/-0.02)
Ld: [0, 1] Ambiguous Link (packing) densely at the interface Interface Stable 0.15 (+/-0.17)
nBSA: [0, 1] Ambiguous Size of the interface relative to the whole complex Interface Variable 0.07 (+/-0.05)
Fintres: [0, 1] Ambiguous Size of the interface relative to the whole complex Interface Variable 0.19 (+/-0.13)
CPscore: [0, 1] Positive inter-residue contacts preference Interface Variable 0.35 (+/-0.17)
CPM: [0, 1] Positive Likelihood of (Sc,EC) given nBSA compared to native Interface Variable 0.80 (+/-0.18)
rTs: [0, 1] Negative Energetic Stability of the whole molecule All_Atom Stable 0.22 (+/-0.21)
Isc: [0, 1] Negative Binding energy Interface Variable 0.20 (+/-0.33)
Erep: [0, 1] Negative Unfavourable Energy due to atomic Clashes All_Atom Stable 0.41 (+/-0.15)
Etmr: [0, 1] Negative Energetic Stability without condiderating clashes All_Atom Stable 0.28 (+/-0.16)
ProQ2: [0, 1] Positive Overall Structural Quality All_Atom Variable 0.80 (+/-0.12)
---------------------------------------------------------------------------------------------------------------------------------
ProQDock [0, 1] Positive The predicted DockQ score, Basu and Wallner, 2016, https://doi.org/10.1371/journal.pone.0161879
DockQ Statistics on CAPRI data:
0 < DockQ < 0.23 - Incorrect
0.23 <= DockQ < 0.49 - Acceptable quality
0.49 <= DockQ < 0.80 - Medium quality
DockQ >= 0.80 - High quality
=================================================================================================================================
EC=0.443
Sc=0.767
rGb=0.07170
Ld=0.14815
nBSA=0.115345
Fintres=0.286517
CPscore=0.490153
CPM=0.94593
rTs=0.04831
Isc=0.00000
Erep=0.34970
Etmr=0.09889
ProQ=0.845
================The ProQDock Score ====================
ProQDock: 1.10501
Features: /tmp/x_bjowa/ProQDock/TESTPDB/1u07.SVMfeatures
ProQDock score: /tmp/x_bjowa/lProQDock/TESTPDB/1u07.ProQDock
Raw Rosetta terms: /tmp/x_bjowa/ProQDock/TESTPDB/1u07.Rterms
Finding correct protein–protein docking models using ProQDock
Sankar Basu and Björn Wallner*
Bioinformatics (2016) 32 (12): i262-i270.
doi: 10.1093/bioinformatics/btw257
The article is avialable online here: http://bioinformatics.oxfordjournals.org/content/32/12/i262.abstract
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent