Empty Output folder - Vina Docking

[jennaj]

Hello @victor_igbokwe

Thanks for posting back the history and the screenshots! I can’t see everything due to the sharing on the history, but that’s Ok. With your descriptions, I think I spot the problem.

On the form and down in the Help section, these two statements fit together:

Select ligands (SDF format with multiple ligands or PDBQT format with single ligand).

If using PDBQT format for the ligands only a single ligand can be specified. If using SDF you can include multiple ligands and those ligands are converted to individual PDBQT format files using openbabel as the first step of tool execution.

Then your comments

@victor_igbokwe

For clarity, my ligand input file in the subject matter job was a single multi-entry SDF file containing up to 695k ligands. In the parameter, I set the ligand file to be split into its components before the Vina Docking. Do you think the error with the empty output folder could be due to the size and number of the ligand input file?

It looks like the Prepare ligands for docking tool also failed when attempting to split up the SDF file. This is a strong clue about what is going wrong with Vina Docking with that same multi-ligand file (not split). I can’t see the job logs from that tool but you could double check them for some more clues.

Vina Docking error: the message is indicating a problem when attempting to parse out the first ligand. In the details, the job is failing at the step where the tool is applying in the xxy size/center and seed parameters but my guess is the tool was attempting to iterate over those parameters when it read in the upstream longer list of ligands, and that is what triggered the problem.

Argument list too long

Why this didn’t fail is just a corner case – any tool can fail or not and still produce unexpected outputs – it just depends on how much error trapping the original tool author included. Galaxy will inherit anything internal to a scientific tool since our wrappers are around the whole package. My guess is that the tool would fail this way even if run directly, outside of Galaxy. You may be able to find discussion online about this.

What to try

Split inputs

• Prepare your inputs with Prepare ligands for docking OR Prepare ligand
• Input the PDBQT collection to VINA Docking
• Each ligand will have a separate job, and those outputs will be written to a collection.

Or, combined inputs

• Run QA/QC with Prepare ligands for docking OR Prepare ligand (to confirm the format can be read by these tools)
• Input the original multi-ligand to VINA Docking. You may need to test to see how many it can handle at once based on any particular search space parameters you apply (coordinates, but also exhaustiveness and others).
• You could put these parameters into separate file(s) and a simple workflow to run all of this in a batch matrix.

Now, I see some failures already for the prepare ligand steps. I think you’ll need to solve this data content problem first, before running the latter tool, to get around the softer “failure” points. That’s why I am recommending using a splitter as a QC step. Then, if the splitting tool can read the data, Vina Docking tool can also likely also read it.

Hope this helps!