Forum Replies Created
- Posts
Hi Christian. Thanks for all your interest in AutoGrow. Unfortunately, the ideal settings will depend very much on the initial population of ligands, both in terms of numbers and available chemical moieties where optimizing fragments can be added. I’d recommend trying a few different settings to see which one gives the best results. Take care.
Hi Xiaoqing. Much thanks for your interest in AutoGrow. I suspect you’re getting this error because you are using AutoGrow settings that are too restrictive. AutoGrow appears to be unable to generate any compounds given those settings. It could also be that your initial starting compound(s) do not have any chemical moieties compatible with AutoGrow’s reaction library. This posting could also be helpful: https://durrantlab.pitt.edu/forums/topic/not-creating-ranked-ligand-file-during-lead-optimization-in-generation-0/
Hope this answer helps. ~Jacob
Hi Patricio. Please forgive my delay in responding. I believe AutoGrow actually redocks the compounds with every generation by default. I suspect the file you’re seeing is the predocked ligand. I’d recommend redocking that compound back into the receptor using AutoDock Vina or some other program, independent of AutoGrow. I hope this answer helps. ~Jacob
Hi Edward. Much thanks for your interest in Gypsum-DL. If I recall correctly, Gypsum prioritize the variants based on their predicted energies, according to RDKit. But there’s also a random component to the variant selection. If you’re not seeing some variants in the generated libraries that you think should be there, you might try allowing a larger number of variants per compound. I hope this helps. Take care.
Hi Jared. Much thanks for your question, and sorry for my delayed response. BlendMol is actually specifically designed not to change the coordinate frame. In my experience, many other ways of exporting molecular meshes do recenter the protein before export. I suspect whatever program you’re using to generate the ply file is the one that’s resetting the frame. Hope this answer helps. Take care.
Hi George. BlendMol uses VMD to generate meshes "under the hood." You can also use VMD directly to set up your scene, and VMD does include a ball-and-sticks representation. You can then save the VMD scene as a visualization state file (. vmd extension) that you can load into Blender via BlendMol. Hope this answer helps.
Hi Christian. Sorry I can’t be more helpful. For what it’s worth, the RDKit library can load PDB files. Maybe it can also handle the PDBQT format. Take care.
Hi Sofea. Much thanks for your interest in Gypsum-DL. ikk6 and I spent some time today trying to get to the bottom of the problem. We just released version 1.1.9, which should avoid errors like those you’re seeing.
Gypsum-DL is primarily designed to work with SMI (SMILES) files. Internally, it uses RDKit to convert SDF files to SMI files. If RDKit can’t process an SDF entry, Gypsum-DL won’t be able to either. Some of your SDF entries cause RDKit problems. I think it’s because there are no charges on some nitrogen atoms. For example, this entry:
CNP0017802 OpenBabel04082123302D 11 11 0 0 0 0 0 0 0 0999 V2000 2.5980 0.0000 0.0000 O 0 0 0 0 0 0 0 0 0 0 0 0 1.7320 -0.5000 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 1.7320 -1.5000 0.0000 O 0 0 0 0 0 0 0 0 0 0 0 0 0.8660 0.0000 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 0.0000 -0.5000 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 -0.9135 -0.0932 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 -1.5826 -0.8364 0.0000 N 0 0 0 0 0 0 0 0 0 0 0 0 -1.0826 -1.7024 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 -0.1045 -1.4945 0.0000 N 0 0 0 0 0 0 0 0 0 0 0 0 0.6386 -2.1636 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 -2.5771 -0.7318 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 1 2 2 0 0 0 0 2 3 1 0 0 0 0 2 4 1 0 0 0 0 4 5 1 0 0 0 0 5 6 2 0 0 0 0 5 9 1 0 0 0 0 6 7 1 0 0 0 0 7 8 1 0 0 0 0 7 11 1 0 0 0 0 8 9 2 0 0 0 0 9 10 1 0 0 0 0 M END > <coconut_id> CNP0017802 > <SMILES> O=C(O)CC1=CN(C=N1C)C $$$$With the latest update, Gypsum-DL will now warn you that it’s skipping an entry like this, rather than throwing an error.
As a side note, if you just use a SMI file, Gypsum won’t get RDKit involved at this step, so you can avoid the problem entirely. You could create an SMI file using open babel like this:
obabel -isdf ligandFile.sdf -osmi > mysmi.smiMuch thanks for bringing this bug to our attention! Take care.
Hi Pan. Much thanks for your interest in DeepFrag. Linking fragments and adding fragments may seem similar from a human perspective, but they are really two fundamentally different tasks. DeepFrag will only perform well on the task it’s trained to perform. It has no way of extrapolating to a different task.
We are very interested in fragment linking, though, and have looked into this question some. We may publish a DeepFrag-like tool for fragment linking in the future.
Take care,
Jacob
I’ve run into the very problem you’re describing. Unfortunately, VMD exports state files as a single mesh. I found that it’s possible to separate them out using Blender’s Mesh->Separate->Selection command. Basically, go into Edit mode, unselect all the vertices, select the ones you want to separate, whether by Material or Select Linked, and then run the Mesh->Separate->Selection command. The selected vertices are then moved to a new object. It’s not exactly a streamlined workflow, but it does work for me most the time.
Another option is to save separate VMD state files, each with a separate representation, and then to import them separately. In some cases that might be more straightforward.
I hope this answer helps!
Hi Rocío. I think you should be fine including the FAD molecule as if it were part of the receptor. I’m assuming that you’re using AutoGrow with AutoDock Vina. My understanding it that Vina can handle receptors that include cofactors: http://plato.cgl.ucsf.edu/pipermail/chimera-users/2014-October/010387.html
Take care.
Hi Christian. Sounds like an interesting project! The PDBQT format includes only polar hydrogen atoms. If I’m not mistaken, the original AutoDock docking program only considers polar hydrogens, which is probably the motivation. My understanding is that AutoDock Vina doesn’t even use the polar hydrogen atoms explicitly, beyond using them to initially assign atom types to the heteroatoms. But I think the PDBQT format retains the polar hydrogens for historical reasons and for AutoDock4 compatibility.
The problem with using Open Babel to reprotonate the docked ligands is that it will essentially undo the protonation states that Gypsum-DL assigns. Really what you need to do is only add back only the non-polar hydrogen atoms (e.g., ones bound to carbon atoms). I’m not aware of a program that can do that, unfortunately. It would probably take some Python coding, using Pybel or RDKit.
Best of luck.
Hi Yutao. Much thanks for your interest in Webina. Really this question applies to AutoDock Vina too. I’ve occasionally seen slightly positive scores in some of my virtual screens. These typically correspond to poses that are not correct. It could be, for example, that there are some clashes between the atoms of the posed ligand and one or more atoms of the protein receptor. You might try expanding the size of your docking box or perhaps increasing the exhaustiveness parameter. Best of luck with your project!
