HELP information for QCS
QCS runs on the command-line on all Unix-compatible platforms.
The source code is available under the GNU GPL license.
It requires gcc >= 3.3.5 and python >= 2.3
Source codes are available here.
You can download it to your local UNIX/LINUX machine by a command like:
Go to the directory that contains QCS_update.tar.gz and install QCS by the following commands:
gzip -d QCS_update.tar.gz
tar -xvf QCS_update.tar
Use command like this to get the QCS (Quality Control Score) for your model):
python getQCS.py [options]
|-new||Required for new targets; Specify if the [real_pdb_file] has not been used before.|
|-r [real_pdb_file]||Required; Input PDB file of real structure [real_pdb_file]|
|-m [model_pdb_file]||Required; Input PDB file for predicted model [model_pdb_file]|
|-opt [w1] [w2] [w3] [w4] [w5] [w6]||Optional; Specify if other sets of weight (as integer or float number) is prefered. From the left to the right
the weight of length score, position score, angle score, interaction score, handedness score and contact score.|
|-h||Optional; Print a help message, and overwrite all other options.|
Final results will be in the same directory as the input [real_pdb_file] and the name is: [real_pdb_file]_vs_[model_pdb_file].final. In this file, the weight and score of each QCS component is
listed. The total socre (weighted average of the QCS component scores) and the final_QCS (normalized total score after comparison with random models) are provided too.
Requirement of the inputs:
The program has stringent requirement of the input files to allow correct "residue matching" between models and targets. It is recommended to pre-parse the PDB files to the most standard PDB
0. Only the "ATOM" records are needed for the computation. (But presence of other information is OK too)
1. Both [real_pdb_file] and [model_pdb_file] need to be in standard PDB format.
2. Both [real_pdb_file] and [model_pdb_file] need to contain only a single chain.
3. [real_pdb_file] and [model_pdb_file] should be the real structure and predicted model for the same protein sequence (but missing residues are allowed) and for a pair of corresponding residues,
their identifiers (6th column) in the PDB files need to match
4. The program assumes the residue identifiers (6th column in the PDB file) to increase residue by residue as going from N-terminus to the C-terminus, and it is better to avoid any "insertion
code" for a residue identifier.
In the "QCS_update" package, under "example/" directory, we have an example file of [real_pdb_file]: "T287.pdb" and examples of [model_pdb_file]: "T287TS020_1", "T287TS092_1" and "T287TS302_1".
You can test QCS using them by the following commands, and the results (T287.pdb_vs_T287TS*.final) will be in the "example/" directory as well.
1. for a target that has not been processed by the program in current directory with equal weights of QCS component scores:
python getQCS.py -r example/T287.pdb -m example/T287TS020_1 -new
2. for a target that has been processed by the program in current directory (obmit the target processing steps to save time when the same target is used to compare with multiple models in the
same directory), with equal weights of QCS components:
python getQCS.py -r example/T287.pdb -m example/T287TS092_1
3. for a target that has not been processed by the program in current directory with customerized weights of QCS component scores:
python getQCS.py -r example/T287.pdb -m example/T287TS302_1 -opt 1 1 2 1 2 2