def calc_contact_order(chimera: Chimera = None, filename: str = None, diss_cutoff: int = 8):
"""
The contact order of a protein is a measure of the locality of the inter-amino acid contacts in the
native folded state. It is computed as the average seqeuence distance between residues that form contacts
below a threshold in the folded protein divided by the total length of the protein"
:param chimera: A Chimera object with n residues.
:param filename: path to a pdb file
:param diss_cutoff: The maximum distance in Armstrong between two residues to be in contact, default 8 Angstroms
:return: the contact order (%)
"""
if chimera and filename:
raise ValueError("Only a Chimera object or the path to a pdb file must be specified")
if not chimera and not filename:
raise ValueError("At least a Chimera object or the path to a pdb file must be specified")
if filename:
chimera = Chimera(filename=filename)
chimera.renumberResidues()
metr = MetricSelfDistance("protein and noh", groupsel="residue", metric="contacts", threshold=diss_cutoff,
pbc=False)
a = metr.project(chimera)
mapping = metr.getMapping(chimera)
matrix, _, _ = contactVecToMatrix(a[0], mapping.atomIndexes)
triang = np.triu(matrix)
idx1, idx2 = np.where(triang)
total_contacts = len(idx1)
total_residues = chimera.numResidues
summation = np.sum(idx2 - idx1)
co = 1 / (total_contacts * total_residues) * summation
print(f"Contact order is {co*100} %")
return co * 100
def getCrystalCO(crystal):
crystalCO = (MetricSelfDistance("protein and name CA",
metric="contacts",
threshold=10,
pbc=False).project(crystal).flatten())
longCO = getLongContacts(crystal)
return crystalCO & longCO
def getCrystalCO(crystal):
crystalCO = MetricSelfDistance('protein and name CA',
metric='contacts',
threshold=10,
pbc=False).project(crystal).flatten()
longCO = getLongContacts(crystal)
return crystalCO & longCO
def getLongContacts(crystal, long=8):
crystalMap = MetricSelfDistance("protein and name CA",
metric="contacts",
threshold=10,
pbc=False).getMapping(crystal)
indexes = np.vstack(crystalMap.atomIndexes.values)
return crystal.resid[indexes[:, 1]] - crystal.resid[indexes[:,
0]] > long
def calc_dist_matrix(chimera: Chimera = None, filename: str = None, selection: str = 'residue', type='contacts',
plot=False):
"""
Returns a matrix of C-alpha distances for a given pdb
:param chimera: A Chimera object with n residues.
:param filename: path to a pdb file
:param selection: How to compute the distance. 'residue' (the closest two
:param type: between contacts (contact map when distances are below 8 armstrongs) or distances atoms between two residues) or 'alpha' distance of the alpha carbons.
:param plot: whether to plot the distance matrix. Default is False
:return: matrix. np.array. An n by n distance matrix.
"""
if chimera and filename:
raise ValueError("Only a Chimera object or the path to a pdb file must be specified")
if not chimera and not filename:
raise ValueError("At least a Chimera object or the path to a pdb file must be specified")
if filename:
chimera = Chimera(filename=filename)
if selection == 'residue':
metr = MetricSelfDistance("protein", groupsel="residue", metric="distances", pbc=False)
mapping = metr.getMapping(chimera)
a = metr.project(chimera)
matrix, _, _ = contactVecToMatrix(a[0], mapping.atomIndexes)
elif selection == 'alpha':
metr = MetricSelfDistance("protein and name CA", metric="distances", pbc=False)
a = metr.project(chimera)
mapping = metr.getMapping(chimera)
matrix, _, _ = contactVecToMatrix(a, mapping.atomIndexes)
else:
raise ValueError("Specify a selection type: 'residue' or 'atom'")
if type == "contacts":
matrix = matrix < 8
elif type != "contacts" and type != "distances":
raise ValueError("Please select contact type between 'contacts' or distances")
if plot:
fig = plt.figure(figsize=(12, 12))
ax = fig.add_subplot(111)
cmap = 'binary'
cax = ax.imshow(matrix, cmap=matplotlib.cm.get_cmap(cmap), interpolation='nearest', origin="lower")
if type == 'distances':
cmap = 'gist_rainbow'
cax = ax.imshow(matrix, cmap=matplotlib.cm.get_cmap(cmap), interpolation='nearest', origin="lower")
cbar = fig.colorbar(cax, cmap=matplotlib.cm.get_cmap(cmap))
plt.xlabel('xlabel', fontsize=24)
plt.ylabel('ylabel', fontsize=24)
plt.xticks(fontsize=22)
plt.yticks(fontsize=22)
plt.xlabel("Residue index")
plt.ylabel("Residue index")
return matrix
md.projection = MetricDistance('protein and name CA', 'resname BEN and noh')
# md.goalprojection = MetricRmsd(Molecule(htmd.home() + '/data/adaptive/generators/1/structure.pdb'),
# 'protein and name CA')
md.goalfunction = rmsdgoal
# md.app = LocalGPUQueue()
# md.run()
# Some real testing now
from moleculekit.projections.metricsecondarystructure import MetricSecondaryStructure
from moleculekit.projections.metricdistance import MetricSelfDistance
import numpy as np
os.chdir(path.join(home(), 'data', 'test-adaptive'))
goalProjectionDict = {'ss': MetricSecondaryStructure(),
'contacts': MetricSelfDistance('protein and name CA', metric='contacts', threshold=10),
'ss_contacts': [MetricSecondaryStructure(),
MetricSelfDistance('protein and name CA', metric='contacts', threshold=10)]}
def getLongContacts(crystal, long=8):
crystalMap = MetricSelfDistance('protein and name CA', metric='contacts', threshold=10, pbc=False).getMapping(
crystal)
indexes = np.vstack(crystalMap.atomIndexes.values)
return crystal.resid[indexes[:, 1]] - crystal.resid[indexes[:, 0]] > long
def getCrystalSS(crystal):
return MetricSecondaryStructure().project(crystal)[0].flatten()
def getCrystalCO(crystal):
crystalCO = MetricSelfDistance('protein and name CA', metric='contacts', threshold=10, pbc=False).project(
crystal).flatten()
return datatica
if __name__ == '__main__':
from htmd.simlist import simlist
from glob import glob
from moleculekit.projections.metricdistance import MetricSelfDistance
from htmd.home import home
from os.path import join
testfolder = home(dataDir='villin')
sims = simlist(glob(join(testfolder, '*', '')),
join(testfolder, 'filtered.pdb'))
met = Metric(sims[0:2])
met.set(MetricSelfDistance('protein and name CA'))
data = met.project()
data.fstep = 0.1
tica = TICA(data, 2, dimensions=range(2, 10))
datatica = tica.project(2)
tica5 = TICA(data, 0.2, units='ns', dimensions=range(2, 10))
datatica5 = tica5.project(2)
expected = [[3.69098878, -0.33862674, 0.85779184],
[3.77816105, -0.31887317, 0.87724227],
[3.83537507, -0.11878026, 0.65236956]]
assert np.allclose(np.abs(datatica.trajectories[0].projection[-3:, -3:]),
np.abs(np.array(expected, dtype=np.float32)),
rtol=0,
atol=0.01)
assert np.allclose(np.abs(datatica5.trajectories[0].projection[-3:, -3:]),
# md.app = LocalGPUQueue()
# md.run()
# Some real testing now
from moleculekit.projections.metricsecondarystructure import (
MetricSecondaryStructure, )
from moleculekit.projections.metricdistance import MetricSelfDistance
os.chdir(path.join(home(), "data", "test-adaptive"))
goalProjectionDict = {
"ss":
MetricSecondaryStructure(),
"contacts":
MetricSelfDistance("protein and name CA",
metric="contacts",
threshold=10),
"ss_contacts": [
MetricSecondaryStructure(),
MetricSelfDistance("protein and name CA",
metric="contacts",
threshold=10),
],
}
def getLongContacts(crystal, long=8):
crystalMap = MetricSelfDistance("protein and name CA",
metric="contacts",
threshold=10,
pbc=False).getMapping(crystal)
indexes = np.vstack(crystalMap.atomIndexes.values)
return datatica
if __name__ == "__main__":
from htmd.simlist import simlist
from glob import glob
from moleculekit.projections.metricdistance import MetricSelfDistance
from htmd.home import home
from os.path import join
testfolder = home(dataDir="villin")
sims = simlist(glob(join(testfolder, "*", "")),
join(testfolder, "filtered.pdb"))
met = Metric(sims[0:2])
met.set(MetricSelfDistance("protein and name CA"))
data = met.project()
data.fstep = 0.1
tica = TICA(data, 2, dimensions=range(2, 10))
datatica = tica.project(2)
tica5 = TICA(data, 0.2, units="ns", dimensions=range(2, 10))
datatica5 = tica5.project(2)
expected = [
[3.69098878, -0.33862674, 0.85779184],
[3.77816105, -0.31887317, 0.87724227],
[3.83537507, -0.11878026, 0.65236956],
]
assert np.allclose(
np.abs(datatica.trajectories[0].projection[-3:, -3:]),
np.abs(np.array(expected, dtype=np.float32)),
