News

KNIME nodes V2.7 released

A new release of workflow components for the KNIME™ environment is now available for Flare™ and Spark™.

In the Cresset node repository (Figure 1 - left), you will find a collection of new Flare KNIME nodes including advanced ligand-based methods validated through years of use in Forge™ (now integrated into Flare) as well as structure-based functionality. You can:

  • Build quantitative and qualitative Structure-Activity Relationship (QSAR) models using the ‘Flare Build’ nodes
  • Score your new molecule designs against QSAR models (built using Flare or the ‘Flare Build’ nodes) with the ’Field Score’ nodes
  • Generate reliable alignments for multiple ligands using Cresset’s patented ligand comparison method with the ‘Flare Align’ node
  • Rapidly dock and score non-covalent and covalent ligands using the new ‘Flare Docking’ node

Nodes for viewing, reading and writing Flare projects and viewing results in Flare are also available (Figure 1 – right).

Flare_KNIME_new-nodes

Figure 1. Left: New Flare nodes are available to build quantitative and qualitative SAR models, score new molecules against the created QSAR models, generate sensible alignments and docking poses for multiple ligands. Right: nodes for viewing, reading and writing Flare projects and viewing results in Flare are also available.

New nodes for building QSAR models and scoring compounds

This release includes new Flare nodes to build qualitative and quantitative models of SAR.

  • ‘Flare Build Field QSAR’ (Figure 1) implements Flare’s Field QSAR method to build models which provide a global view of your SAR data. Field QSAR models give you both prediction of activity and interpretation of results, enabling you to understand the reason a particular compound is highly active or inactive, to inform the design of the next generation of compounds.

Flare_KNIME_build-field-qsar-nodes

Figure 2. The ‘Flare Build Field QSAR’ node can be used to build quantitative Field QSAR models for your ligands. Basic and advanced options give you full control on model building.

  • ’Flare Build kNN’ uses the robust, and well validated, machine learning k-Nearest Neighbors method for building predictive quantitative SAR regression and classification models.
  • ‘Flare Field Descriptors’ can be used to generate Cresset 3D field descriptors, modeling the electrostatic properties and shape of aligned ligand, which you can use to build quantitative SAR models with a variety of KNIME nodes.
  • ‘Field Score Field QSAR/kNN’: once a robust, predictive quantitative SAR model is calculated in Flare or using the ‘Build’ nodes, the ‘Field Score Field QSAR/kNN’ nodes use the model to calculate the activity of new molecule designs, aiding the prioritization of their synthesis, enabling you to focus on the most promising compounds.
  • ‘Flare Build Activity Atlas™’ and ‘Flare Score Activity Atlas’: the Activity Atlas method is useful to summarize the SAR for your ligand series in qualitative 3D maps, enabling you to focus on important SAR signals and to identify unexplored regions to solve key problems in the project. This method is available in the ‘Flare Build Activity Atlas’ node. The ‘Flare Score Activity Atlas’ node uses an Activity Atlas model calculated with the ‘Build’ node or in Flare to calculate the novelty of new molecule designs, to see whether they will take you into new regions, or only send you to places that you have been before.

New nodes for generating sensible alignments and poses for your ligands

‘Flare Align’ (Figure 3) aligns and scores molecules to one or more reference ligands using Cresset ligand alignment method, generating biologically relevant comparisons of active ligands. When used on a congeneric series, this method can help in library design by giving a rationale for the prioritization of compounds for synthesis. The ‘substructure’ alignment option is useful to generate inputs into QSAR model building.

Flare_KNIME_align-node

Figure 3. Use the ‘Flare Align’ node to align and score your ligands against one or more reference molecules.

‘Flare Docking’ (Figure 4) runs docking and scoring experiments using the Lead Finder™ docking algorithm implemented in Flare. Docking in Flare combines a genetic algorithm search with local optimization procedures, which make it efficient in providing promising docking solutions. Three different scoring functions are available for the accurate prediction of 3D docked ligand poses, protein-ligand binding energy and rank-ordering of active and inactive compounds in virtual screening experiments. The ‘Flare Docking’ node also supports covalent docking and offers template docking as an option to generate enhanced docking results for congeneric lead series.

Flare_KNIME_docking-node

Figure 4. Run docking and scoring experiments for covalent and non-covalent ligands using the ‘Flare Docking’ node.

New and enhanced example workflows

See the Flare and Spark nodes in action in the new and enhanced Cresset KNIME example workflows (Figure 5). These simple workflows encompass each of the Cresset nodes showcasing their functionality and providing practical examples of usage in your integrated workflows.

Flare_KNIME_new-enhanced-workflows

Figure 5. New and enhanced Cresset KNIME example workflows are freely available for download.

Try the Cresset KNIME nodes on your project

Contact Cresset support to get these new components free of charge.

If you’re not currently using Cresset desktop solutions you can request an evaluation to try them on your project.

Try Cresset solutions on your project