The DNAproDB processing pipeline is built as a collection of python scripts that take as input a PDB entry (in mmCIF or PDB file format) and produce a single JSON file as output containing all information about the DNA and protein structural characteristics and the interactions between protein residues and nucleotides in the complex. Additional annotations from several other databases are used as well. A schematic of the pipeline is shown below.
It is preferred to use the mmCIF file format, as the header info is parsed and additonal annotations can be extracted from it, which would be missing for the PDB file format.
Clone this repository and install python dependencies via pip
git clone https://github.com/timkartar/dnaprodb.git
cd dnaprodb
conda env create -f environment.yml
conda activate dnaprodb
In addition to the python packages listed in python-requirements.txt, dnaprodb relies on a variety of command line tools that are invoked by the processing scripts. The following must be compiled or made executable and accessible from the system path:
- reduce https://github.com/rlabduke/reduce
- hbplus (source available in
/share/hbplus.tar.gz) - Curves 5.3 (source available in
/share/curves5.3_edited.tar.gz) - 3DNA (source available in
x3dna-v2.3.tar.gz) - pdb2pqr (source available in
pdb2pqr-3.5.2.zip) - dssr (available as linux binary in
/share/x3dna-dssr) - dssp (available as linux binary in
/share/dssp) - SNAP (available as linux binary in
/share/x3dna-snap) - msms (available as linux binary in
/share/msms)
To run the pipeline, simply invoke the main script processStructure.py from the command line as such:
./processStructure.py <FILE_NAME> --clean
the output will be a single file containing all the DNAproDB output data in a file called <FILE_NAME_PREFIX>.json.