Get volume of a single tunnel in several similar proteins and limit tunnel length outwards

[GuhLelouch]

Hi all,

First of all, I apologize if this is not the proper place to ask usage questions.

Now, to the questions:

1. I have several structurally similar proteins (wild-types) and I am interested in measuring the volume of a particular tunnel in these proteins in order to evaluate their capacity to accommodate a certain ligand. What is the recommended way to automate this task? I was thinking about aligning all structures to one with the ligand, save the structures with a copy of the ligand, and then use dpocket on a .txt input file containing the line "/workdir/protein_lig_complex.pdb LIG" for each complex. When I tested this, the first pocket identified was indeed the one of interest, but dozens of other pockets that I am not interested in were also identified. That leads to my second question:
2. When focusing on one ligand with dpocket, is there a way to turn off the identification of other pockets?
3. Now, I'm mostly interested in the volume on the dead end of the tunnel because this is where I hypothesize that the specificity for a longer ligand is controlled. I want the tunnel to stop being computed at a fixed position before it reaches the exit of the tunnel. That is because the volume variation in this region may introduce noise to the analysis. Hence, my question is: how to limit how "far" the tunnel will be computed towards the exit, supposing that it starts on its dead end? I've tried using for instance
   fpocket -f /workdir/complex.pdb -P 89::A.148::A.156::A.157::A.247::A.288::A.318::A.319::A.348::A.350::A.378::A.64::B
   but I get

! No pocket to refine! (argument NULL: (nil)).
no pockets found

4. I've also tried
   fpocket -f /workdir/complex.pdb -r 393:LIG:H
   And it worked. It identified only the pocket of interest. However, it also identified several regions outside the tunnel of interest as a pocket (the ligand is long and is partly solvent-exposed. Can I limit this using an option during the run, or will -r always have priority over those options? I've tried reducing -D, but it didn't change the results.

Thank you in advance!

P.S.: I'm running fpocket on Docker on Windows.

[pschmidtke]

if it's a tiny tunnel you probably need to tweak a bit the minimum alpha sphere parameter.
I'd do this with mdpocket and using an input list of pdb files. Get a first run, extract the tunnel and then run the descriptor calculations with mdpocket as well