rcsbscan is a biopython-based tool for large-scale RCSB protein structure database screening with a user-defined 3D structural template (reference structure).
The amino acid sequence of the template is completely ignored in the scan process, only atomwise RMSD of the reference structure with residue tuples by all backbone atoms is the criteria for the structural hit. Residue tuple is the sequence of residues with defined 3D structure of the same length as the reference structural template provided. For example, 9-residue protein have four six-reside tuples to be checked for structural hit:
1-ABCDEFGHI-9 -- scanned structure from the database
XXXXXX |
|XXXXXX |
| XXXXXX|
| XXXXXX -- reference structure
All rmsd hits are printed to the terminal with match description. The structures found are optionally saved to the separate PDB file for future analysis.
The RCSB protein databease should be downloaded locally and preferably to the hard drive of the computational server. Please follow the instructions provided by RCSB on the following link: https://www.rcsb.org/docs/programmatic-access/batch-downloads-with-shell-script
rcsbscan.py ARGS
Scan database by sequential atomwise fit of the provided reference
structure to every substructute
Positional arguments:
struct Structure(s) in pdb, mmcif or mmtf format, optionally gzipped.
Directory or a list of directories with protein structures is
allowed with -r flag (see below). To scan full RCSB database,
first download it locally and privide it with an additional
-r option
Optional arguments:
-h, --help show this help message and exit
-p, --print-header Print header of the PDB which is scanned
--ref-structure REF_STRUCTURE
Reference structure, in PDB, MMCIF of MMTF formats
--ref-model REF_MODEL
Model number in the reference structure, e.g. '0'
--ref-chain REF_CHAIN
Chain in the reference structure, e.g. 'A'
--ref-residues REF_RESIDUES
Residue range in the reference structure, e.g. '26-33'
--ref-atoms REF_ATOMS
Atoms in reference structure.
By default, only four atoms per residue are considured:
N, CA, C, O
-w, --pdb-warnings Show structure parsing warnings
-v, --verbose Verbose output
-r, --recursive Recursive search of structures in folders
--max-rms MAX_RMS Maximum RMSD to print [ default 1.0 A ]
--water WATER Water molecule to be included to scan (residue number)
--water-max-rms WATER_MAX_RMS
Max rms for water match [ default 2.0 A ]
--save-pdb-hits SAVE_PDB_HITS
Save pdb hits to file
--renumber-pdb Renumber residues in the output pdb hits
--xray-res XRAY_RES Maximal resolution of X-ray structures to scan
--xray-only Scan only X-ray structures
--ncpu NCPU Number of CPU to use [ All available CPU's are used by default ]
Examples:
Ussuming that RCSB clone is downloadad to the folder ~/RCSB/pdb/:
rcsbscan.py ~/RCSB/pdb/m0/pdb1m0k.ent.gz --ref-structure ./examples/alpha-helix_A10.pdb
-- scans the 1M0K structure of bacteriorhodopsin at 1.43 A resolution
( (2002) J Mol Biol 321: 715-726 )
atomwise and prints all structural hits with an alpha-helix template provided in the example file.
The output will find numerous alpha helices in the structure of bacteriorhodopsin:
save pdb hits: None
REF_4FIT: ALPHA-HELIX_A10 model= 0, chain= seq= 1 AAAAAAAAAA 10 atoms=N,CA,C,O fit_atoms=40 max_rms=1.0000
Start fit scan
Start the pool of 8 CPU (of 8 available)
struct list: ['/home/maxim/RCSB/pdb/m0/pdb1m0k.ent.gz'] (trancated at 10 structures)
nstruct: 1
Prepare arguments for 1 structures
Start map_async
map_async submitted 1 tasks to Pool of 8 cpu
RMSD_HIT: 1M0K XRay 1.43A model= 0, chain= A size= 222 hit= 8 PEWIWLALGT 17 rms= 0.8595
RMSD_HIT: 1M0K XRay 1.43A model= 0, chain= A size= 222 hit= 9 EWIWLALGTA 18 rms= 0.3121
RMSD_HIT: 1M0K XRay 1.43A model= 0, chain= A size= 222 hit= 10 WIWLALGTAL 19 rms= 0.2138
RMSD_HIT: 1M0K XRay 1.43A model= 0, chain= A size= 222 hit= 11 IWLALGTALM 20 rms= 0.2530
RMSD_HIT: 1M0K XRay 1.43A model= 0, chain= A size= 222 hit= 12 WLALGTALMG 21 rms= 0.2440
....
1M0K XRay 1.43A model= 1, chain= A size= 222 hit= 81 ARYADWLFTT 90 rms= 0.8996
1M0K XRay 1.43A model= 0, chain= A size= 222 hit= 82 RYADWLFTTP 91 rms= 0.9623
1M0K XRay 1.43A model= 1, chain= A size= 222 hit= 82 RYADWLFTTP 91 rms= 0.9623
Overall statistics:
fitscan statistics after 3 sec:
1 files with structutes
0 files skipped
1 files scanned
2 structures in all models/chains
228 hits
406 tuples of 10 residues superimposed and rms of atoms N,CA,C,O evaluated
0 errors
Evaluation time: 3.82
No filename was provided to store the results found
Use the command-line argimens '--save-pdb-hits FILE' to save all matches in PDB format