CaviTracer tool for detecting channels, tunnels, pores

[karolamik13]

CaviTracer, created by Eryk Trzcinski, as his MSc work under my supervision.
Additional installation of open3D package is required:
pip install open3d

Test functions:
from prody import *

1. Single PDB:
   p = parsePDB('1tqn.pdb')
   atoms = p.select("protein")
   channels, surface = calcChannels(atoms, output_path='channels_output.pdb')
   showCavities(surface)

To create 3D model to display it in ProDy
vmd_path = '/usr/local/bin/vmd'
model = getVmdModel(vmd_path, atoms)

To display channels (all or one by one):
showChannels(channels, surface=surface, model=model)
showChannels(channels, model=model)
getChannelParameters(channels)
showCavities(surface, show_surface=True)

Save all channels independently as PDB:
output = 'chanels_test'
channels, surface = calcChannels(atoms, output, separate=True)

Display channel 1:

showChannels(channels[1], model)

selected_channels = [channels[1], channels[8]]
showChannels(selected_channels, model)

selected_channels = channels[1:4]
lengths, bottlenecks, volumes = getChannelParameters(selected_channels)
selected_channels_atoms = getChannelAtoms(selected_channels)

atoms_protein = getChannelAtoms(channels, atoms)

Extracting residues that are forming channel[1]:
atoms_protein = getChannelAtoms(channels[1], atoms)
len(atoms_protein.getResnames())

distance = 4
residues = atoms_protein.select('same residue as exwithin '+str(distance)+' of resname FIL')
residues.select('name CA').getResnames(), residues.select('name CA').getResnums()
list(zip(residues.select('name CA').getResnames(), residues.select('name CA').getResnums()))

2. Multiple frames:
   p2 = parsePDB('LXmulti.pdb')
   channels2, surfaces2 = calcChannelsMultipleFrames(p2)

To display particular channel:
channels2[4][1]

model2 = getVmdModel(vmd_path, p2)
showChannels(channels2[4][1], model2)
showChannels(channels2[4][2], model2)

3. Trajectories:
   PDBfile = 'L_traj.pdb'
   DCDfile = 'L_traj.dcd'
   atoms = parsePDB(PDBfile)
   dcd = Trajectory(DCDfile)
   dcd.link(atoms)
   dcd.setCoords(atoms)

channels3, surfaces3 = calcChannelsMultipleFrames(atoms, dcd)
all_channels, all_surfaces=calcChannelsMultipleFrames(atoms, dcd, output_path = 'channels_dcd', separate=True)

[jamesmkrieger]

@jamesmkrieger did you get a chance to test this?

I don't think so really

[karolamik13]

At the current stage, a few more changes are needed. CaviTracer can be omitted in the upcoming ProDy release.

[jamesmkrieger]

Someone still needs to go through and check this in detail, which I haven't managed to do yet

At any rate, it would be good to have unit tests please

[karolamik13]

Hi James,
I plan to do it in this week. I will prepare tutorial in Jupyter Notebook file for other people who would be checking/testing CaviTracer functions.

[jamesmkrieger]

Sounds good!

[jamesmkrieger]

Maybe make that an option

[karolamik13]

Maybe make that an option
If we were to have hydrogen atoms included, we should add various sphere radius for different atoms (hydrogens are crucial here being much smaller), and that will slow down a lot of predictions of the channels/tunnels. Lack of hydrogen atoms will not change much the outcome (I consulted with an expert in that field). Later, I can think about including a variety of spheres, but for now, this is the best option.

[jamesmkrieger]

Maybe make that an option
If we were to have hydrogen atoms included, we should add various sphere radius for different atoms (hydrogens are crucial here being much smaller), and that will slow down a lot of predictions of the channels/tunnels. Lack of hydrogen atoms will not change much the outcome (I consulted with an expert in that field). Later, I can think about including a variety of spheres, but for now, this is the best option.

ok, sounds good