3D-e-Chem Nodes for KNIME

About the nodes

The 3D-e-Chem nodes have been developed as part of the 3D-e-Chem project by Vrije Universiteit Amsterdam, Radboudumc Nijmegen and Netherlands eScience Center. The nodes complement existing cheminformatics and bioinformatics nodes to enable the efficient exploitation of structural and pharmacological protein-ligand interaction data from proteome-wide databases, as well as customized information systems focused on e.g. G Protein-Coupled Receptors (GPCRdb) and protein kinases (KLIFS). The 3D-e-Chem KNIME node toolbox provides building blocks for the design of flexible computer-aided drug discovery workflows, including ligand-based metabolism prediction (SyGMA), pharmacophore-based (align-it) and shape-based (shape-it) alignment of molecules, protein sequence analyses (ss-TEA), and structure-based protein binding site comparison and bioisosteric replacement for ligand design (KRIPO), molecular docking (PLANTS) and the visualization and alignment of protein-ligand complexes (Molviewer) and pharmacophores (Pharmacohore). The open source, freely available Virtual Machine, 3D-e-Chem-VM facilitates the efficient use of pre-configured 3D-e-Chem tools and other resources.

Current 3D-e-Chem nodes have been grouped as follows:

• GPCRDB, Extraction of sequence, structural, protein-ligand interaction, and site-directed mutagenesis information of GPCRs from the GPCRdb database.
• KLIFS, Extraction of sequence, structural, and protein-ligand interaction information of kinases from the KLIFS database.
• KRIPOdb, nodes to identify KRIPO (Key Representation of Interaction in Pockets) pharmacophore based similarities between protein binding sites and corresponding ligand substructures.
• Molviewer, Web-based 3D molecule viewer of ligands, proteins and pharmacophores.
• SyGMa, Systematic generation of potential phase 1 and phase 2 metabolite structures. The SyGMa node is a thin wrapper around the SyGMa Python library
• Pharmacophore aligner, readers and writers
• PLANTS, to set up, run, and analyse molecular docking simulations.
• Silicos-it, sets of nodes for pharmacophore-based (align-it) and shape-based (shape-it) alignment of molecules, property-based filtering of chemical databases (filter-it), and chemical scaffold analysis (strip-it).
• ss-TEA score, Entropy-based identification of receptor-specific ligand binding residues
• Modified Tanimoto distance measure that can be selected in the "Bit Vector Distance" node of the KNIME Distance matrix extension.

Examples

The GitHub repository of each node has a simple example workflow, including:

• Chemdb4VS workflow for the evaluation and optimization of virtual screening strategies
• GPCRdb example workflow for the extraction and combination of structural, sequence, and mutation data for specific receptors from GPCRdb
• KLIFS example workflow for the integrated analysis of structural kinase-ligand interactions, kinase binding sites, and kinase ligand features for a specific kinase in KLIFS.
• KRIPO example workflow for the identification of protein-ligand binding site similarities and possibilities for bioisosteric replacements by structure-based ligand design.
• Ligand-based cross-reactivity prediction
• Scaffold replacements for kinase ligand design
• Sequence-based ligand repurposing within a protein family
• Structure-based bioactivity data mapping of kinase inhibitors workflow.
• Structure-based GPCR-kinase cross-reactivity prediction workflow for off-target identification, ligand repurposing, or the discovery of ligands with a desired GPCR-kinase polypharmacology profile
• SyGMa example workflow for metabolite prediction.

Several example workflows have been described in the 3D-e-Chem application note, and a list of 3D-e-Chem workflows can be found here.

About the 3D-e-Chem project

The 3D-e-Chem project, involving the Vrije Universiteit Amsterdam, Radboudumc Nijmegen and Netherlands eScience Center, develops technologies to improve the integration of ligand and protein data for structure-based prediction of protein-ligand selectivity and polypharmacology. The project uses the KNIME Analytics Platform to integrate different structural cheminformatics and bioinformatics technologies and datasets.

Source Code

The source code can be accessed at https://github.com/3D-e-Chem. Each node has its own repository, for example, https://github.com/3D-e-Chem/knime-gpcrdb contains the source code for the GPCRDB nodes.

License

The 3D-e-Chem nodes are released under GNU GPL version 3 license.