search

ccsb-scripps / AutoDock-GPU

AutoDock for GPUs and other accelerators
https://ccsb.scripps.edu/autodock
GNU General Public License v2.0
391 stars 107 forks source link

How to Disable Desolvation Parameters in AutoDock Scoring Function during Docking? #253

Open samdani1593 opened 8 months ago

samdani1593 commented 8 months ago

Hello,

I am currently working on a docking project using AutoDock, and I would like to switch off the desolvation parameters in the scoring function. I have reviewed the AutoDock documentation, but I couldn't find clear instructions on how to achieve this.

The Autodock version trying to implement: 4.2.6

If there are code modifications required, could you please guide me through the changes?

rwxayheee commented 8 months ago

Hi @samdani1593 Have you tried to include a custom parameter file in your DPF? According to the AutoDock4 Manual pp. 39, you could point AutoDock4 to a modified parameter file. According to the documentation here, you could set FE_coeff_desolv to 0 in the modified parameter file.

diogomart commented 8 months ago

There's two categories of interacting atom pairs:

  1. between static (receptor) and movable atoms (ligand and flexible sidechain)
  2. between movable atoms

For the first, which is arguably the most important one, desolvation needs to be disabled in autogrid, by the means that @rwxayheee explained: passing a modified parameter file with FE_coeff_desolv set to zero. This will disable dissolvation between the static atoms (receptor) and the moving atoms (ligand and flexible sidechains).

The second category of interactions, which are pairs of movable atoms, can't be disabled in AutoDock-GPU. DPF files are not fully supported, and the parameter_file keyword is ignored. It can be disabled in Vina with the command line option --weight_ad4_dsolv. The autogrid modification is still required for vina, because vina reads autogrid maps when using the AD4 scoring function.

atillack commented 8 months ago

Although AD-GPUs atom types are currently fixed along with the corresponding coefficients (corresponding to AD4.1_bound.dat from AD4.2.6) changing them based on user input is something we did think about but haven't found the time to actually do ...

Now with the first user feedback this may be needed, I will turn this into a feature request.