"""This module defines input and output functions for NMD format.
.. _nmd-format:
NMD Format
-------------------------------------------------------------------------------
Description
^^^^^^^^^^^
NMD files (extension :file:`.nmd`) are plain text files that contain at
least normal mode and system coordinate data.
NMD files can be visualized using :ref:`nmwiz`. ProDy functions
:func:`.writeNMD` and :func:`.parseNMD` can be used to read and write NMD
files.
Data fields
^^^^^^^^^^^
Data fields in bold face are required. All data arrays and lists must be in a
single line and items must be separated by one or more space characters.
**coordinates**: system coordinates as a list of decimal numbers
Coordinate array is the most important line in an NMD file. All mode array
lengths must match the length of the coordinate array. Also, number of atoms
in the system is deduced from the length of the coordinate array.
::
coordinates 27.552 4.354 23.629 24.179 4.807 21.907 ...
**mode**: normal mode array as a list of decimal numbers
Optionally, mode index and a scaling factor may be provided
in the same line as a mode array. Both of these must precede the mode array.
Providing a scaling factor enables relative scaling of the mode arrows and
the amplitude of the fluctuations in animations. For NMA, scaling factors
may be chosen to be the square-root of the inverse-eigenvalue associated
with the mode. Analogously, for PCA data, scaling factor would be the
square-root of the eigenvalue.
If a mode line contains numbers preceding the mode array, they are evaluated
based on their type. If an integer is encountered, it is considered the mode
index. If a decimal number is encountered, it is considered the scaling
factor. Scaling factor may be the square-root of the inverse eigenvalue
if data is from an elastic network model, or the square-root of the
eigenvalue if data is from an essential dynamics (or principal component)
analysis.
For example, all of the following lines are valid. The first line contains
mode index and scaling factor. Second and third lines contain mode index or
scaling factor. Last line contains only the mode array.
::
mode 1 2.37 0.039 0.009 0.058 0.038 -0.011 0.052 ...
mode 1 0.039 0.009 0.058 0.038 -0.011 0.052 ...
mode 2.37 0.039 0.009 0.058 0.038 -0.011 0.052 ...
mode 0.039 0.009 0.058 0.038 -0.011 0.052 0.043 ...
*name*: name of the model
The length of all following data fields must be equal to the number of atoms in
the system. NMWiz uses such data when writing a temporary PDB files for
loading coordinate data into VMD.
*atomnames*: list of atom names
If not provided, all atom names are set to "CA".
*resnames*: list of residue names
If not provided, all residue names are set to "GLY".
*chainids*: list of chain identifiers
If not provided, all chain identifiers are set to "A".
*resids*: list of residue numbers
If not provided, residue numbers are started from 1 and incremented by one
for each atom.
*bfactors*: list of experimental beta-factors
If not provided, all beta-factors are set to zero.
Beta-factors can be used to color the protein representation.
NMD files may contain additional lines. Only lines that start with one of the
above field names are evaluated by NMWiz.
Autoload Trick
^^^^^^^^^^^^^^
By adding a special line in an NMD file, file content can be automatically
loaded into VMD at startup. The first line calls a NMWiz function to load the
file itself (:file:`xyzeros.nmd`).
::
nmwiz_load xyzeros.nmd
coordinates 0 0 0 0 0 0 ...
mode 0.039 0.009 0.058 0.038 -0.011 0.052 ...
mode -0.045 -0.096 -0.009 -0.040 -0.076 -0.010 ...
mode 0.007 -0.044 0.080 0.015 -0.037 0.062 ...
In this case, VMD must be started from the command line by typing
:program:`vmd -e xyzeros.nmd`."""
__all__ = ['parseNMD', 'writeNMD', 'pathVMD', 'getVMDpath', 'setVMDpath',
'viewNMDinVMD']
import os
from os.path import abspath, join, split, splitext
import numpy as np
from prody import LOGGER, SETTINGS, PY3K
from prody.atomic import AtomGroup
from prody.utilities import openFile, isExecutable, which, PLATFORM, addext
from .nma import NMA
from .anm import ANM
from .gnm import GNM, ZERO
from .pca import PCA
from .mode import Vector, Mode
from .modeset import ModeSet
[docs]def pathVMD(*path):
"""Returns VMD path, or set it to be a user specified *path*."""
if not path:
path = SETTINGS.get('vmd', None)
if isExecutable(path):
return path
else:
LOGGER.warning('VMD path is not set by user, looking for it.')
vmdbin = None
vmddir = None
if PLATFORM == 'Windows':
if PY3K:
import winreg
else:
import _winreg as winreg # PY3K: OK
for vmdversion in ('1.8.7', '1.9', '1.9.1'):
try:
key = winreg.OpenKey(winreg.HKEY_LOCAL_MACHINE,
'Software\\University of Illinois\\VMD\\' +
vmdversion)
vmddir = winreg.QueryValueEx(key, 'VMDDIR')[0]
vmdbin = join(vmddir, 'vmd.exe')
except:
pass
try:
key = winreg.OpenKey(winreg.HKEY_LOCAL_MACHINE,
'Software\\WOW6432node\\University of Illinois\\VMD\\' +
vmdversion)
vmddir = winreg.QueryValueEx(key, 'VMDDIR')[0]
vmdbin = join(vmddir, 'vmd.exe')
except:
pass
else:
vmdbin = which('vmd')
if False:
pipe = os.popen('which vmd')
vmdbin = pipe.next().strip()
vmdfile = open(vmdbin)
for line in vmdfile:
if line.startswith('defaultvmddir='):
vmddir = line.split('=')[1].replace('"', '')
break
vmdfile.close()
if isExecutable(vmdbin):
setVMDpath(vmdbin)
return vmdbin
elif len(path) == 1:
path = path[0]
if isExecutable(path):
SETTINGS['vmd'] = path
SETTINGS.save()
LOGGER.info("VMD path is set to '{0}'.".format(path))
else:
raise OSError('{0} is not executable.'.format(str(path)))
else:
raise ValueError('specify a single path string')
[docs]def getVMDpath():
"""Deprecated for removal in v1.5, use :func:`pathVMD` instead."""
return pathVMD()
[docs]def setVMDpath(path):
"""Deprecated for removal in v1.5, use :func:`pathVMD` instead."""
return pathVMD(path)
NMD_LABEL_MAP = {
'atomnames': 'name',
'resnames': 'resname',
'resnums': 'resnum',
'resids': 'resnum',
'chainids': 'chain',
'bfactors': 'beta',
'segnames': 'segment',
'segments': 'segment',
}
[docs]def parseNMD(filename, type=NMA):
"""Returns :class:`.NMA` and :class:`.AtomGroup` instances storing data
parsed from *filename* in :file:`.nmd` format. Type should be :class:`.NMA`
or a subclass such as :class:`.PCA`, :class:`.ANM`, or :class:`.GNM`."""
if isinstance(type, str):
type = type.upper().strip()
if 'ANM' in type:
type = ANM
elif 'GNM' in type:
type = GNM
elif 'PCA' in type or 'EDA' in type:
type = PCA
elif type == 'NMA':
type = NMA
else:
type = None
if not issubclass(type, NMA):
raise TypeError('type must be NMA, ANM, GNM, or PCA')
atomic = {}
atomic.update([(label, None) for label in NMD_LABEL_MAP])
atomic['coordinates'] = None
atomic['name'] = None
modes = []
with open(filename) as nmd:
for i, line in enumerate(nmd):
try:
label, data = line.split(None, 1)
except ValueError:
try:
label = line.strip()
data = line[len(label):-1]
LOGGER.warn('Blank data associated with label {0}'
'.'.format(repr(label)))
except ValueError:
LOGGER.warn('Could not process data associated with label {0}'
'.'.format(repr(label)))
pass
if label == 'mode':
modes.append((i + 1, data))
elif label in atomic:
if atomic[label] is None:
atomic[label] = (i + 1, data)
else:
LOGGER.warn('Data label {0} is found more than once in '
'{1}.'.format(repr(label), repr(filename)))
for i, label in enumerate(NMD_LABEL_MAP.keys()):
if atomic[label] is None:
atomic[label] = (i + 1, None)
LOGGER.warn('The data label {0} was not found in '
'{1}.'.format(repr(label), repr(filename)))
name = atomic.pop('name', '')[1].strip() or splitext(split(filename)[1])[0]
ag = AtomGroup(name)
dof = None
n_atoms = None
line, coords = atomic.pop('coordinates', None)
if coords is not None:
coords = np.fromstring(coords, dtype=float, sep=' ')
dof = coords.shape[0]
if dof % 3 != 0:
LOGGER.warn('Coordinate data in {0} at line {1} is corrupt '
'and will be omitted.'.format(repr(filename), line))
else:
n_atoms = dof // 3
coords = coords.reshape((n_atoms, 3))
ag.setCoords(coords)
from prody.atomic import ATOMIC_FIELDS
for label, data in atomic.items(): # PY3K: OK
if data is None:
continue
line, data = data
if data is None:
continue
if len(data) == n_atoms:
data = ['']*n_atoms
else:
data = data.split()
if n_atoms is None:
n_atoms = len(data)
dof = n_atoms * 3
elif len(data) != n_atoms:
LOGGER.warn('Data with label {0} in {1} at line {2} is '
'corrupt, expected {3} values, parsed {4}.'.format(
repr(label), repr(filename), line, n_atoms, len(data)))
continue
label = NMD_LABEL_MAP[label]
data = np.array(data, dtype=ATOMIC_FIELDS[label].dtype)
ag.setData(label, data)
if not modes:
return None, ag
length = len(modes[0][1].split())
is3d = length > n_atoms + 2
if dof is None:
dof = length - (length % 3)
elif not is3d: # GNM
dof = n_atoms
array = np.zeros((dof, len(modes)))
less = 0
eigvals = []
count = 0
for i, (line, mode) in enumerate(modes):
mode = np.fromstring(mode, dtype=float, sep=' ')
diff = len(mode) - dof
if diff < 0 or diff > 2:
LOGGER.warn('Mode data in {0} at line {1} is corrupt.'
.format(repr(filename), line))
continue
array[:, i - less] = mode[diff:]
count += 1
eigvals.append(mode[:diff])
if count == 0:
return None, ag
try:
eigvals = np.array(eigvals, dtype=float)
except TypeError:
LOGGER.warn('Failed to parse eigenvalues from {0}.'
.format(repr(filename)))
if eigvals.shape[1] > 2:
LOGGER.warn('Failed to parse eigenvalues from {0}.'
.format(repr(filename)))
eigvals = None
elif eigvals.shape[1] == 1:
if np.all(eigvals % 1 == 0):
LOGGER.warn('Failed to parse eigenvalues from {0}.'
.format(repr(filename)))
eigvals = None
else:
eigvals = eigvals.flatten() ** 2
else:
eigvals = eigvals[:, 1] ** 2
nma = type(name)
if type != PCA:
eigvals = 1./eigvals
if count != array.shape[1]:
array = array[:, :count].copy()
nma.setEigens(array, eigvals)
return nma, ag
[docs]def writeNMD(filename, modes, atoms):
"""Returns *filename* that contains *modes* and *atoms* data in NMD format
described in :ref:`nmd-format`. :file:`.nmd` extension is appended to
filename, if it does not have an extension.
.. note::
#. This function skips modes with zero eigenvalues.
#. If a :class:`.Vector` instance is given, it will be normalized
before it is written. It's length before normalization will be
written as the scaling factor of the vector."""
if not '.nmd' in filename:
filename += '.nmd'
if not isinstance(modes, (NMA, ModeSet, Mode, Vector)):
raise TypeError('modes must be NMA, ModeSet, Mode, or Vector, '
'not {0}'.format(type(modes)))
if modes.numAtoms() != atoms.numAtoms():
raise Exception('number of atoms do not match')
out = openFile(addext(filename, '.nmd'), 'w')
#out.write('#!{0} -e\n'.format(VMDPATH))
out.write('nmwiz_load {0}\n'.format(abspath(filename)))
name = modes.getTitle()
name = name.replace(' ', '_').replace('.', '_')
if not name.replace('_', '').isalnum() or len(name) > 30:
name = str(atoms)
name = name.replace(' ', '_').replace('.', '_')
if not name.replace('_', '').isalnum() or len(name) > 30:
name = splitext(split(filename)[1])[0]
out.write('name {0}\n'.format(name))
try:
coords = atoms.getCoords()
except:
raise ValueError('coordinates could not be retrieved from atoms')
if coords is None:
raise ValueError('atom coordinates are not set')
try:
data = atoms.getNames()
if data is not None:
out.write('atomnames {0}\n'.format(' '.join(data)))
except:
pass
try:
data = atoms.getResnames()
if data is not None:
out.write('resnames {0}\n'.format(' '.join(data)))
except:
pass
try:
data = atoms.getResnums()
if data is not None:
out.write('resids ')
data.tofile(out, ' ')
out.write('\n')
except:
pass
try:
data = atoms.getChids()
if data is not None:
out.write('chainids {0}\n'.format(' '.join(data)))
except:
pass
try:
data = atoms.getSegnames()
if data is not None:
out.write('segnames {0}\n'.format(' '.join(data)))
except:
pass
try:
data = atoms.getBetas()
if data is not None:
out.write('bfactors ')
data.tofile(out, ' ', '%.2f')
out.write('\n')
except:
pass
format = '{0:.3f}'.format
out.write('coordinates ')
coords.tofile(out, ' ', '%.3f')
out.write('\n')
count = 0
if isinstance(modes, Vector):
out.write('mode 1 {0:.2f} '.format(abs(modes)))
modes.getNormed()._getArray().tofile(out, ' ', '%.3f')
out.write('\n')
count += 1
else:
if isinstance(modes, Mode):
modes = [modes]
for mode in modes:
if mode.getEigval() < ZERO:
continue
out.write('mode {0} {1:.2f} '.format(
mode.getIndex()+1, mode.getVariance()**0.5))
arr = mode._getArray().tofile(out, ' ', '%.3f')
out.write('\n')
count += 1
if count == 0:
LOGGER.warning('No normal mode data was written. '
'Given modes might have 0 eigenvalues.')
out.close()
return filename
[docs]def viewNMDinVMD(filename):
"""Start VMD in the current Python session and load NMD data."""
vmd = pathVMD()
if vmd:
os.system('{0} -e {1}'.format(vmd, abspath(filename)))
