Library to interact with Kripo fragment, fingerprint and similarity data files.

KRIPO stands for Key Representation of Interaction in POckets, see reference for more information.

• Pocket, binding site of the ligand in the protein of a crystal structure
• Fragment, part of the ligand
• Subpocket, part of the protein pocket which binds with the fragment
• Fingerprint, fingerprint of structure-based pharmacophore of subpocket
• Similarity matrix, similarities between all fingerprint pairs calculated using the modified tanimoto similarity index
• Kripo fragment identifier, used as identifier for fragment, subpocket and fingerprint

Requirements:

• RDKit, http://www.rdkit.org/docs/Install.html, to read SDF files and generate smile strings from molecules
• pip, version 8.0.0 or greater, for wheel support
• git, to clone kripodb repository during installation

pip install -U setuptools
pip install numpy
pip install git+https://github.com/3D-e-Chem/kripodb.git

To see available commands

kripodb --help

Commands to create all data files see update documentation.

Command to find fragments most similar to 3kxm_K74_frag1 fragment.

kripodb similar sim_all.h5 3kxm_K74_frag1 --cutoff 0.45

Commands to create similarity matrix see update documentation.

An example data set included in the data/ directory of this repo. See data/README.md for more information.

All fragments based on GPCR proteins compared with all proteins in PDB.

• kripo.gpcrandhits.sqlite - Fragments sqlite database
• kripo.gpcr.h5 - HDF5 file with similarity matrix

The data set has been published at

All fragments form all proteins-ligand complexes in PDB compared with all.

• Fragments sqlite database - Download from http://3d-e-chem.vu-compmedchem.nl/kripodb/fragments.sqlite
• Similarity matrix - Can be queried on webservice at http://3d-e-chem.vu-compmedchem.nl/kripodb. For build instructions see http://kripodb.readthedocs.io/en/latest/data-update.html
• Fragment fingerprints - See http://kripodb.readthedocs.io/en/latest/data-update.html for instructions how to convert to a similarity matrix

A data set with PDB entries till 23 December 2015 has been published at

The Knime-KripoDB-example.zip file is an example workflow showing how to use KripoDB python package inside KNIME (http://www.knime.org). It can be run by importing it into KNIME. Make sure the Python used by KNIME is the same as the Python with kripodb package installed.

The https://github.com/3D-e-Chem/knime-kripodb repo adds KripoDB code templates to KNIME.

Install the development deps with:

pip install -r requirements.txt

docker build -t 3dechem/kripodb .

Show the kripodb help with

docker run --rm 3dechem/kripodb kripodb --help

To calculate the mean bit density of the fingerprints in the fingerprints.sqlite file in the current working directory use following command.

docker run --rm -u $UID -v $PWD:/data 3dechem/kripodb kripodb meanbitdensity /data/fingerprints.sqlite

The Kripo data files can be queried using a web service.

Start webservice with:

kripodb serve data/similarities.h5 data/fragments.sqlite data/pharmacophores.h5

It will print the urls for the swagger spec and UI.

Note! The webservice returns a limited amount of results. To get all results use local files.

On http://3d-e-chem.vu-compmedchem.nl/kripodb/ui/ there is a KripoDB webservice with the full PDB fragment all vs all matrix.

API and data update pipeline documentation can be found at http://kripodb.readthedocs.io/en/latest/.

KRIPO – a structure-based pharmacophores approach explains polypharmacological effects; Tina Ritschel, Tom JJ Schirris, and Frans GM Russel; J Cheminform. 2014; 6(Suppl 1): O26; Published online 2014 Mar 11; http://dx.doi.org/10.1186/1758-2946-6-S1-O26