Losing Zinc metal in Macromolecule when converting to PDBQT

[sargis]

The best way to handle this currently is to edit pdbqt and assign charges to Mg atoms there:
viewtopic.php?f=11&t=376

I'll see if there are other ways to do this more effectively. Thanks.

[M Indarte]

Sargis- It works fine. No need to invest more time on this.

I did this:
Navigator--> Autodock Molecules--> select macromolecule-->right click-->EDIT-->I go to the pdb text file and change the charges.

Thanks again!
Martin.

PS: I have really interesting files to show you. They might give you several ideas for future PyRx release... Is the sargis@ email that appears in PyRx if you want to report a bug or suggestion in use?

[sargis]

Thanks Martin! Please email at sargis @ scripps.edu to report any bug or suggestion. I look forward to hearing new ways to improve PyRx. Thanks!

[GryderArt]

Thank you both for your input, I've learned quite a lot in the last week. I'm now running a series of experiments with different charges on the Metal in the binding site using various macromolecules and ligands that our research group is interested in. Early results seem to suggest that the charged Zn allows for better docking (better BA, correct orientations, etc).

Cheers!
~Berkley

[M Indarte]

Oh, we are all learning!

Anyways, I have changed the charges of my Mg like this:

Original PDBQT (got this report in the python shell: Unable to assign MAP type to atom Mg Sorry, there are no Gasteiger parameters available for atom 3egz_Mg++_nolig_fromPyRX: :MG 1:Mg)

HETATM 1643 HN4 C65 130 13.432 22.126 83.663 1.00 0.00 0.157 HD
HETATM 1644 HN4 C65 130 14.423 20.857 82.978 1.00 0.00 0.157 HD
HETATM 1645 Mg MG 1 29.485 31.434 69.018 1.00 0.00 0.000 Mg
TER 1646 MG 1

Then I have changed the last column to:

HETATM 1643 HN4 C65 130 13.432 22.126 83.663 1.00 0.00 0.157 HD
HETATM 1644 HN4 C65 130 14.423 20.857 82.978 1.00 0.00 0.157 HD
HETATM 1645 Mg MG 1 29.485 31.434 69.018 1.00 0.00 2.000 Mg
TER 1646 MG 1

I do docking with the uncharged and charged macromolecule and I get the same results. Am I changing charges in the correct way? Are you guys doing it like this?

Cool beans,

Martin.

[sargis]

HETATM 1644 HN4 C65 130 14.423 20.857 82.978 1.00 0.00 0.157 HD
HETATM 1645 Mg MG 1 29.485 31.434 69.018 1.00 0.00 2.000 Mg

I do docking with the uncharged and charged macromolecule and I get the same results. Am I changing charges in the correct way? Are you guys doing it like this?

This seems to be OK. Are you using AutoDock 4 or Vina? Vina ignores charges in pdbqt, that's why you'll get the same result in both cases.

[M Indarte]

OK. Yes, I am using VINA. I know that partial charges are pretty much ignored, was not sure about the formal 2+ charge. I am planning to do more docking with the same system but manually changing Mg2+ to Zn2+ to see if there is an improvement or change in scores and also geometries. I guess that the charge would not really matter then, only the volume of the substituted atom?

M.

[GryderArt]

This seems to be OK. Are you using AutoDock 4 or Vina? Vina ignores charges in pdbqt, that's why you'll get the same result in both cases.

I'm not sure what else the difference could be for mine... I changed the charges on my Zn the same way MI changed the charges on his Mg and I got totally different results using VINA. I'm going to rerun this to make sure there wasn't a difference in grid size or something else that caused this difference.

Thanks!

[GryderArt]

I did notice that the conf.txt showed slightly different

Code: Select all

(Zn = 0)
receptor = HDAC8_H_add_mod3.pdbqt
center_x = 63.1381781721
center_y = 65.7104666329
center_z = -0.556966828136

(Zn = 2)
receptor = HDAC8_H_add_mod4.pdbqt
center_x = 63.5961
center_y = 65.5476
center_z = -0.3963

Could that account for large differences in results?

[sargis]

I did two quick runs with two different charges on Zn and got different results too. The reason for getting different results, AFAIK, is not the charge on Zn but the fact that Vina uses stochastic optimization. In other words, you might get different results even with the same input. Here are my results from two different runs

Zn.png (45.23 KiB) Viewed 2457 times

The last number in Ligand table shows the charge on Zn (0 or 2) and the number before _Zn is the run number (1 or 2). There was no difference in the input between two runs, but results were still different.