Schrödinger KNIME Extensions allow scientists to protoype, validate, automate, and deploy multi-step workflows. In this installment of KNIME Questions and Answers, Schrödinger’s KNIME Extensions Associate Product Manager Jean-Christophe Mozziconacci talks about how to use KNIME to refine and analyze the results of conformer searches, and how to incorporate Python scripts into KNIME workflows.
Q. There are certain operations that I regularly perform in Maestro, but I don’t think there are any KNIME nodes for those tasks. Is it still possible to create a KNIME workflow that includes those tasks?
A. If a task can be performed in Maestro, it can be performed in a KNIME workflow using a node called Run Maestro Command (Schrödinger > Tools > Scripting > Run Maestro Command). A trivial example of this might be to render certain residues in a specific color. When you configure the node, simply paste the desired Maestro commands into the “Command line(s)” field.
If you’re not sure which commands are used to execute the task in Maestro, try opening Maestro’s Command Script Editor panel (Maestro > Command Script Editor in the Maestro menu bar, or Ctrl+E). As you perform the task you’re interested in, you’ll see that Maestro logs the corresponding commands in this interface. Once you know which commands to use, simply copy and paste the relevant part of the command history.
--
Q. What can I do if there are no extensions capable of performing the task I’m interested in? Can I include Python, Java or Perl scripts in a KNIME workflow?
A. Even if you can’t find a node that does exactly what you want, it is still probably possible to perform that task within a KNIME workflow.
You can run Python scripts – including Python scripts that take advantage of Schrödinger APIs – using one of our Python script nodes (Schrödinger > Tools > Scripting > Python Script). You can also use KNIME’s JPython nodes for more basic tasks (Python > JPython Function or Python > JPython Script).
Command line Python scripts are easily modified to run within a Python script node. KNIME stores information in the input table and the output data table. Hence the Python script must specify the columns to extract from a KNIME table, and it must also specify the columns to create in the node’s output table. The extracted columns contain input for the script, and the newly created columns are used to store results. For more information, see the usage cases described in the node documentation.
Similarly, Java and Perl scripting nodes are also available in the KNIME workbench (Misc > Java Snippet and KNIME Labs > Perl Scripting).
The “Chemistry External Tool” nodes will also allow you to run any shell commands. For a simple example of this, see the workflow called “Molecular Mechanics - Compare ConfSearch Methods.” This node can be used to run any Schrödinger tool through the command line.
There are various versions of the Python script nodes and the Chemistry External Tool nodes. The nodes differ in the number of input and output connectors. Both allow a maximum of 2 input and 2 output connectors.
Feel free to contact us so we can improve an existing node or create a new one.
--
Q. I would like to create an automated process that my coworkers can use to refine ligand conformers using ab initio techniques. How can I do this with KNIME?
A. The example workflow “Conformational Search and Quantum Mechanics Refinement” carries out this exact task, and includes many options specific to individual needs. The workflow can take several structures as input, process them using the method described above, and then calculate the Boltzmann populations as well as the value of the dihedral angle that discriminates between conformers.
These values are then used to compare the conformational spaces before and after QM optimization. A report (Figure 1) is created with the Report Designer functionality developed by KNIME.com, available free of charge (see details on http://knime.org/report-designer) and included in our installers.
Figure 1: Example report created by the “Conformational Search and Quantum Mechanics Refinement” workflow using the KNIME Report Designer.
--