PDBLibrary

Toolset to create a library of .pdb files for the virtual screening for possible binders (or for usage with finDr).

Dependencies

• python3

Installation

Simply clone this repository into a directory of your liking

cd ilikethisdirectory
git clone https://github.com/kcaliban/PDBLibrary.git 

Preparing a library

For this small tutorial we assume that you are in the directory that you have cloned this repository into, if you aren't, cd ilikethisdirectory. You need about 200 GB of space for this process, a lot more if you want to not just extract but also slice helices from the Protein Data Bank.

Getting the latest version of the Protein Data Bank

First, we need to download the latest version of the Protein Data Bank, here into a folder called PDB:

mkdir PDB
cd PDB
wget -r -np -nd ftp://ftp.wwpdb.org/pub/pdb/data/structures/divided/pdb/*

This will leave you with a lot of ent.gz files. These are compressed files, so we have to decompress them (using a for loop since gunzip has a parameter limit).

for i in *.gz; do gunzip $i; done

After this process is done, you need to rename the files.

for i in *.ent; do rename 's/ent$/pdb/g' $i; done

You now have the latest version of the whole Protein Data Bank on your hard drive. Go back to the original directory by cd ...

Extracting all helices

To get an even bigger library, you can extract all helices from this dataset using the Helix Extractor. Make sure you have enough disk space (about 90GB extra) and that you are in the parent directory of PDB and execute the following:

mkdir helices
python3 PDBLibrary/helixextract.py PDB helices

This will take a while. After successful completion, the folder helices will contain all helices extracted from the Protein Data Bank. The files have the following format:

HELIXLENGTH_FILENAME_HELIXNUMBER.pdb

The filenames allow for easier extraction of specified lengths and later identification of the protein.

If you only want the helices of say length 12, you can copy these into a directory called e.g. helices_12 by

mkdir helices_12
for i in helices/12_*; do cp $i helices_12; done

Removing duplicates and files with alternate locations

A lot of helices, especially short ones, reappear in different proteins. Also, some structures contain alternate locations for CA atoms, for which both get interpreted as seperate atoms in programs like Autodock Vina. To make this library ready for tools like Autodock Vina, these files get discarded.

To remove duplicates and files with alternate locations of helices of length 12:

mkdir helices_12_clean
python3 PDBLibrary/duplrem.py helices_12 helices_12_clean

The folder helices_12_clean now contains only distinct helices and none with alternate locations.

Creating even bigger libraries: Slicing helices

To go even further, you can slice helices into smaller helices of specified length(s). This requires a looot of disk space, so make sure you have enough.

To get helices of length 9 out of our prior extracted helices of length 12:

mkdir helices_sliced_9
python3 PDBLibrary/helixslice.py helices_12_clean helices_sliced_9

To get helices of length 8-12 all prior extracted helices:

for i in {8..12}; do mkdir helices_sliced_$i; python3 PDBLibrary/helixslice.py helices helices_sliced_$i $i; done

Included Tools

Helix Extractor

Extracts all helices specified in the header of one or more .pdb files.

To extract all helices from a single file

python3 helixextract.py file

To extract all helices from all files in a folder

python3 helixextract.py inputdir outputdir

Helix Slicer

Slices the helices of .pdb files into parts of specified length using a "sliding window" of one or more .pdb files.

To slice .pdbs of length 5 from a file

python3 helixslice.py file 5

To slice .pdbs of length 10 from all files in a folder

python3 helixslice.py inputdir outputdir 10

Duplicate Remover

Removes duplicate sequences in a pool of .pdb files as well as all .pdb files that contain alternate locations (since these affect AutoDock Vina scores)

Outputdir contains "cleaned" library. Files only get moved and not copied for time efficiency.

python3 duplrem.py inputdir outputdir

License

See LICENSE file