QUELO FAQ
How should I prepare the PDB for my target receptor?
QSimulate offers a protein preparation tool called ProtClean. QUELO also accepts input from most other protein preparation software and will issue a warning with details of what you should fix if there is a problem with the file.
What format file do I need to upload for my ligands?
An SDF file (in 3D V2000 format) is required for all ligands.2D SDF files (e.g., where all “z” coordinates are 0) will not work.
Do I upload my ligands in one concatenated file, or as a series of separate files?
Do ligands need to be already aligned to each other?
No, alignment is performed automatically by the platform.
Can I evaluate multiple rotamers for a ligand?
Yes. Start by enabling “expert mode” under your account tab. Then, after uploading your ligands, you can click on a ligand to bring up a 2D image of it. At the bottom, you will find a “Rotate Bond” section. Enter the atom numbers associated with the start and end of the bond you wish to rotate. Then click “save.” This will generate a second input conformer for the molecule, with atoms connected to the second atom of the specified rotation bond rotated by 180 degrees. Both conformers will be attached to the same vertex (node) of the mutation map. The second conformer will appear in the results tables with a name post-pended by “-alt”.
Does the platform allow you to generate separate protomers for ligands?
Not currently. We will introduce this functionality in the future. At present, the user must choose a protomer state upon input, and that protomer will be retained during the calculation. You can upload multiple protomeric states for a ligand as separate molecules, and the program will treat them as separate nodes in the mutation map and the output tables.
Is there a command-line interface (CLI) or API for QUELO?
Yes, QUELO can be accessed either through the browser-based GUI or else from a CLI. The full suite of functionality is available using either access method.
Do I need to be an FEP expert to run QUELO?
No. One of the strengths of QUELO is that it automates the complex workflow required to run a free energy calculation. Structure preparation, ligand superposition, atom correspondence, mutation map generation, individual job submission, and post-processing are all carried out automatically.
Does QUELO work with nucleic acids?
Yes, QUELO works with both protein and nucleic acid receptors
Can QUELO work with ligand binding that also incorporates co-factors or structural waters?
Yes, you can include additional moieties in the QM region
Does QUELO support metalloproteins?
Yes to both.
How does QUELO treat the solvent?
QUELO uses an explicit water model, TIP3P.
What are the limitations of QUELO FEP in terms of ligand similarity?
The limitations for QUELO are the same as with any implementation of FEP. Ligands must be similar enough to allow for suitable sampling between adjacent lambda states. This requirement is identical for QUELO and classical FEP methods.
Is QUELO a modified version of FEP, such as an endpoint method?
No. QUELO is implemented exactly the same as classical FEP. The only difference is that the ligand and surrounding residues can be treated using QM. Energies and forces are calculated at the QM level for every step of MD sampling. Multiple nanoseconds of MD sampling are performed for each FEP window.
Does QUELO provide error estimates for the values it calculates?
Yes. QUELO determines error bounds using two approaches. For each mutation, a lower error bound is determined from BAR (Bennett Acceptance Ratio) analysis of adjacent distributions. And, overall errors are estimated from a comparison of free energies determined along multiple paths connecting nodes in the mutation map.