def matchAlign(mobile, target, **kwargs):
"""Superpose *mobile* onto *target* based on best matching pair of chains.
.. versionadded:: 0.7.1
This function makes use of :func:`matchChains` for matching chains.
This function returns a tuple that contains the following items:
* *mobile* after it is superposed,
* Matching chain from *mobile* as a :class:`~prody.atomic.AtomMap`
instance,
* Matching chain from *target* as a :class:`~prody.atomic.AtomMap`
instance,
* Percent sequence identity of the match,
* Percent sequence overlap of the match.
"""
match = matchChains(mobile, target, **kwargs)
if not match:
return
match = match[0]
LOGGER.info('RMSD before alignment (A): {0:.2f}'
.format(prody.calcRMSD(match[0], match[1])))
prody.calcTransformation(match[0], match[1]).apply(mobile)
LOGGER.info('RMSD after alignment (A): {0:.2f}'
.format(prody.calcRMSD(match[0], match[1])))
return (mobile,) + match
def doPy(self):
print "calculating ...."
import prody
"""
-----
pip install prody
or
http://prody.csb.pitt.edu/index.html
------
RMSF[i]=sqrt(ave((x[i]-average_x[i])**2))
delta2=(x[i]-average_x[i])**2
B-factor=8*pi**2*RMSF**2/3
Ref: Willis & Pryor, Thermal vibrations in crystallography, Cambridge Univ. Press, 1975
unit: Angstrom**2
"""
tmpcoord=[]
nframes=len(self.coordinates)
ave_coordinate=np.average(self.coordinates,axis=0)
rmsf=np.zeros(len(self.coordinates[0]))
for j in range(nframes):
t=prody.calcTransformation(ave_coordinate,self.coordinates[j])
for i in range(self.natoms):
delta2=(self.coordinates[j][i].dot(t.getRotation())-ave_coordinate[i])
rmsf[i]+=delta2.dot(delta2)
self.cal_bfactor=np.sqrt(rmsf/nframes)**2*8*np.pi**2/3
print "done!!"
return np.sqrt(rmsf/nframes)**2*8*np.pi**2/3
def orient(pdb, selection='all'):
act = pdb.select(selection)
adj = prody.calcCenter(act)
oldcoords = act.getCoords()
newcoords = np.subtract(oldcoords, adj)
nncoords = varimax(newcoords)
trans = prody.calcTransformation(oldcoords, nncoords)
trans.apply(pdb)
return pdb
def compare_pdb_files(file1, file2):
"""Returns the RMSD between two PDB files of the same protein.
Args:
file1 (str): Path to first PDB file.
file2 (str): Path to second PDB file. Must be the same protein as in file1.
Returns:
float: Root Mean Squared Deviation (RMSD) between the two structures.
"""
s1 = pr.parsePDB(file1)
s2 = pr.parsePDB(file2)
transformation = pr.calcTransformation(s1, s2)
s1_aligned = transformation.apply(s1)
return pr.calcRMSD(s1_aligned, s2)
def rmsd(a, b):
"""Return the RMSD between two sets of coordinates."""
t = pr.calcTransformation(a, b)
return pr.calcRMSD(t.apply(a), b)
def compute_transformation(self):
self.transformation = prody.calcTransformation(self.mobile, self.target)
def _do_align(self):
self._transformation = prody.calcTransformation(self._prediction, self._native)
self._transformation.apply(self._prediction)
rmsd = prody.calcRMSD(self._native, self._prediction)
self._align_results = RMSDAlignmentResult(rmsd)
def calcS2(model_list, S2_records, S2_type, fit, fit_range):
"""Returns a dictonary with the average S2 values:
S2_calced[residue] = value"""
# fitting models
reference = model_list[0]
if fit and not PDB_model.is_fitted:
print("Start FITTING")
for i in range(1, len(model_list)):
mobile = model_list[i]
matches = prody.matchChains(reference, mobile)
match = matches[0]
ref_chain = match[0]
mob_chain = match[1]
if fit_range:
weights = np.zeros((len(ref_chain), 1), dtype=np.int)
fit_start, fit_end = fit_range.split('-')
for i in range(int(fit_start) - 1, int(fit_end) - 1):
weights[i] = 1
else:
weights = np.ones((len(ref_chain), 1), dtype=np.int)
t = prody.calcTransformation(mob_chain, ref_chain, weights)
t.apply(mobile)
PDB_model.is_fitted = True
# get NH vectors from models (model_data[] -> vectors{resnum : vector})
model_data = []
s2_pairs = {'N': 'H', 'CA': 'HA'}
for model in model_list:
current_Resindex = 1
has_first, has_second = False, False
vectors = {}
for atom in model:
# why not .getResnum() ???
atom_res = atom.getResindex() + 1
if atom_res != current_Resindex:
current_Resindex = atom_res
has_first, has_second = False, False
if atom_res == current_Resindex:
if atom.getName() == S2_type:
has_second = True
N_coords = Vec_3D(atom.getCoords())
elif atom.getName() == s2_pairs[S2_type]:
has_first = True
H_coords = Vec_3D(atom.getCoords())
if has_first and has_second:
has_first, has_second = False, False
vectors[atom_res] = Vec_3D(N_coords - H_coords).normalize()
model_data.append(vectors)
S2_calced = {}
# iterating over STR records
for resnum in [int(s2rec.resnum) for s2rec in S2_records]:
x2, y2, z2, xy, xz, yz = 0, 0, 0, 0, 0, 0
# iterating over PDB models
for m in model_data:
# coordinates in model at a given resnum
x, y, z = m[resnum].v[0], m[resnum].v[1], m[resnum].v[2]
x2 += x**2
y2 += y**2
z2 += z**2
xy += x * y
xz += x * z
yz += y * z
x2 /= len(model_data)
y2 /= len(model_data)
z2 /= len(model_data)
xy /= len(model_data)
xz /= len(model_data)
yz /= len(model_data)
# S2 calcuation
s2 = 3 / 2.0 * (x2**2 + y2**2 + z2**2 + 2 * xy**2 + 2 * xz**2 +
2 * yz**2) - 0.5
S2_calced[resnum] = s2
return S2_calced
def s2_values(
model_data, calculate_on_models, s2_records, s2_type, fit, fit_range
):
"""Returns a dictionary with the average S2 values:
s2_calced[residue] = value"""
if fit:
reference = model_data.atomgroup[:]
model_data.atomgroup.setACSIndex(0)
prody.alignCoordsets(model_data.atomgroup.calpha)
if fit_range:
for model_num in calculate_on_models:
model_data.atomgroup.setACSIndex(model_num)
mobile = model_data.atomgroup[:]
matches = prody.matchChains(reference, mobile)
match = matches[0]
ref_chain = match[0]
mob_chain = match[1]
weights = np.zeros((len(ref_chain), 1), dtype=np.int)
fit_start, fit_end = fit_range.split("-")
for i in range(int(fit_start) - 1, int(fit_end) - 1):
weights[i] = 1
t = prody.calcTransformation(mob_chain, ref_chain, weights)
t.apply(mobile)
# get NH vectors from models (model_data[] -> vectors{resnum : vector})
vector_data = []
s2_pairs = {"N": "H", "CA": "HA"}
h_coords = None
n_coords = None
for model_num in calculate_on_models:
model_data.atomgroup.setACSIndex(model_num)
current_resindex = 1
has_first, has_second = False, False
vectors = {}
for atom in model_data.atomgroup:
atom_res = atom.getResnum()
if atom_res != current_resindex:
current_resindex = atom_res
has_first, has_second = False, False
if atom_res == current_resindex:
if atom.getName() == s2_type:
has_second = True
n_coords = Vec3D(atom.getCoords())
elif atom.getName() == s2_pairs[s2_type]:
has_first = True
h_coords = Vec3D(atom.getCoords())
if has_first and has_second:
has_first, has_second = False, False
vectors[atom_res] = Vec3D(
n_coords - h_coords
).normalize()
vector_data.append(vectors)
s2_calced = {}
# iterating over STR records
for resnum in [int(s2rec.resnum) for s2rec in s2_records]:
x2, y2, z2, xy, xz, yz = 0, 0, 0, 0, 0, 0
# iterating over PDB models
for m in vector_data:
# coordinates in model at a given resnum
x, y, z = m[resnum].v[0], m[resnum].v[1], m[resnum].v[2]
x2 += x ** 2
y2 += y ** 2
z2 += z ** 2
xy += x * y
xz += x * z
yz += y * z
x2 /= len(vector_data)
y2 /= len(vector_data)
z2 /= len(vector_data)
xy /= len(vector_data)
xz /= len(vector_data)
yz /= len(vector_data)
s2 = (
3
/ 2.0
* (
x2 ** 2
+ y2 ** 2
+ z2 ** 2
+ 2 * xy ** 2
+ 2 * xz ** 2
+ 2 * yz ** 2
)
- 0.5
)
s2_calced[resnum] = s2
return s2_calced
def _do_align(self):
self._transformation = prody.calcTransformation(
self._prediction, self._native)
self._transformation.apply(self._prediction)
rmsd = prody.calcRMSD(self._native, self._prediction)
self._align_results = RMSDAlignmentResult(rmsd)