def gen_cmap(args):
"""Load a trajectory file and calculate the atom contactmap.
Parameters
----------
args: argparse object
the argument options
Returns
-------
cmap_dat: pd.DataFrame, shape = [ N, M]
the contactmap along the time. N is number of frames,
M is N_res * N_res (N_res is number of residues for cmap
calculations. )
"""
# TODO: add a residue based selection module here
# select atoms indices for distance calculations
atoms_selections = args.select.split()
if len(atoms_selections) == 2:
atoms_selections = atoms_selections
else:
atoms_selections = [atoms_selections[0]] * 2
top = mt.load(args.s).topology
atom_grp_a = top.select("name %s" % atoms_selections[0])
atom_grp_b = top.select("name %s" % atoms_selections[1])
# load the trajectory file by chunks iteratively
print("Iterloading xtc trajectory file ...... ")
trajs = gmxcli.read_xtc(xtc=args.f, top=args.s, chunk=1000, stride=int(args.dt/args.ps))
cmap_dat = np.array([])
print("Computing contactmap ...... ")
for i, traj in enumerate(trajs):
contmap = cmap.ContactMap(traj=traj, group_a=atom_grp_a, group_b=atom_grp_b, cutoff=args.cutoff)
contmap.generate_cmap(shape="array", switch=args.switch)
if i == 0:
cmap_dat = contmap.cmap_
else:
cmap_dat = np.concatenate((cmap_dat, contmap.cmap_), axis=0)
cmap_dat = pd.DataFrame(cmap_dat)
return cmap_dat
def iterload_xyz_coordinates(xtcfile, top, chunk, stride, atom_selection="name CA"):
"""
Load a large gromacs xtc trajectory file using mdtraj.iterload method, and extract
the atom xyz coordinates.
Parameters
----------
xtcfile: str, format gromacs xtc
the gromacs xtc trajectory file
top: str, format pdb
the reference pdb file
chunk: int,
number of frames to load per time
stride: int,
skip n frames when loading trajectory file
atom_selection: str,
the atom selection language
Returns
-------
xyz: pandas dataframe, shape = [N, M]
the xyz coordinates dataset,
N is number of samples, or frames
M is n_atoms * 3
"""
trajs = gmxcli.read_xtc(xtc=xtcfile, top=top, chunk=chunk, stride=stride)
xyz = np.array([])
for i, traj in enumerate(trajs):
copca = cmap.CoordinatesXYZ(traj, top, atom_selection)
if i == 0:
xyz = copca.xyz_coordinates()
else:
xyz = np.concatenate((xyz, copca.xyz_coordinates()), axis=0)
xyz = pd.DataFrame(xyz)
return xyz
def gen_dihedrals(args):
"""General dihedral angles from gromacs xtc file
Parameters
----------
args : argparse object,
the arguments options
Returns
-------
di_angles: pd.DataFrame, shape = [ N, M ]
time-series dihedral angles. N is number of frames, M is number of
dihedral angles per frame.
"""
#args = arguments(d=d)
elements = angles.read_index(args.n, angle_type="dihedral")
dih_angles = np.array([])
# load the trajectory file by chunks iteratively
print("Loading xtc trajectory file now ......")
trajs = gmxcli.read_xtc(args.f, args.s, chunk=200, stride=int(args.dt / args.ps))
# calculate the dihedral angles by chunks iteratively
print("Calculating dihedral angles ...... ")
for traj in trajs:
dang = angles.ComputeAngles(traj)
if dih_angles.shape[0] == 0:
dih_angles = dang.get_dihedral_angles(elements)
else:
dih_angles = np.concatenate((dih_angles, dang.get_dihedral_angles(elements)), axis=0)
#if not isinstance(dih_angles, pd.DataFrame):
dih_angles = pd.DataFrame(dih_angles)
dih_angles.index = np.arange(dih_angles.shape[0]) * args.dt
return dih_angles
def run_coord_number():
"""
Examples:
gmx_coordnum.py -f test_traj_dt1ns.xtc -s reference.pdb -rc " " 1 40
-lc " " 50 70 -o coord_result.csv -atomtype all all -byres True
-v True -cutoff 0.5
"""
startTime = datetime.now()
args = arguments()
resides_a, resides_b = [], []
group_a, group_b = [], []
if args.byres:
ndx = index.PdbIndex(args.s, [args.rc[0]], args.rc[1:])
ndx.resid_mapper()
#print(ndx.resid_mapper_)
s, e = ndx.resid_mt_style(args.rc[0], args.rc[1], args.rc[2])
#print(s, e)
resides_a = np.arange(s, e + 1)
ndx = index.PdbIndex(args.s, [args.lc[0]], args.lc[1:])
ndx.resid_mapper()
#print(ndx.resid_mapper_)
s, e = ndx.resid_mt_style(args.lc[0], args.lc[1], args.lc[2])
resides_b = np.arange(s, e + 1)
verbose(args.v,
"X-axis resids : " + " ".join([str(x) for x in resides_a]))
verbose(args.v,
"Y-axis resids : " + " ".join([str(x) for x in resides_b]))
else:
# TODO: define a way to select atom slices
# define the atom indices for receptor
'''ndx = index.PdbIndex(reference=args.s, atomtype=args.atomtype[0],
resSeq=args.rc[1:],
chain=[args.rc[0]])
ndx.prepare_selection()
ndx.res_index()'''
group_a = index.gen_atom_index(pdbin=args.s,
chain=[args.rc[0]],
atomtype=args.atomtype[0],
resSeq=args.rc[1:],
style="mdtraj")
#print(group_a)
# define the atom indices for ligand
'''ndx = index.PdbIndex(reference=args.s, atomtype=args.atomtype[1],
resSeq=args.lc[1:],
chain=[args.lc[0]])
ndx.prepare_selection()
ndx.res_index()'''
group_b = index.gen_atom_index(pdbin=args.s,
chain=[args.lc[0]],
atomtype=args.atomtype[-1],
resSeq=args.lc[1:],
style="mdtraj")
verbose(args.v, "Atom indexing processing completed ......")
results = np.array([])
verbose(args.v, "Loading trajectory xtc file ...... ")
trajs = gmxcli.read_xtc(args.f,
args.s,
chunk=1000,
stride=int(args.dt / args.ps))
verbose(args.v, "Performing calculations ...... ")
for i, traj in enumerate(trajs):
verbose(args.v,
"Generate coordinate number for chunk #%d trajectory " % i)
if args.byres:
coord = cmap.CmapNbyN(traj,
resids_a=resides_a,
resids_b=resides_b,
cutoff=args.cutoff)
coord.contact_nbyn()
if i == 0:
results = coord.contacts_by_res_
else:
results = np.concatenate((results, coord.contacts_by_res_),
axis=0)
else:
coord = cmap.ContactMap(traj, group_a, group_b, args.cutoff)
coord.coord_num()
if i == 0:
results = coord.coord_number_
else:
results = np.concatenate((results, coord.coord_number_),
axis=0)
results = pd.DataFrame(results)
results.index = np.arange(results.shape[0]) * args.dt
verbose(args.v, "Saving results to output file ...... ")
results.to_csv(args.o,
sep=",",
header=False,
index=True,
float_format="%.1f")
print("Total Time Usage: ")
print(datetime.now() - startTime)
def gmxangle():
"""A gromacs g_angle simulator which works the same way as the gromacs tool
Returns
-------
angles: np.ndarray, shape=[N * M ]
the angles, N is number of frames, M is the number of angles per frame
"""
args = arguments()
if os.path.exists(args.f) and os.path.exists(args.n) \
and os.path.exists(args.s):
# prepare index atom slices
ndx = read_index(args.n, args.type)
if args.v:
print("Compute cosine of the angles: ", bool(args.cos))
print("Atom indices: ")
print(ndx)
# load trajectories
trajs = gmxcli.read_xtc(xtc=args.f,
top=args.s,
chunk=1000,
stride=int(args.dt / args.ps))
if args.v:
print("Frame information: ")
for i, traj in enumerate(trajs):
print("Trajectory %3d: %12d frames" % (i, traj.n_frames))
angles = np.array([])
# for each traj chunk, calculate angles, and cat them together
for i, traj in enumerate(trajs):
cangle = ComputeAngles(traj)
if args.v:
print("Progress: %12d " % (i * traj.n_frames))
print(angles.shape)
if angles.shape[0] == 0:
angles = cangle.get_dihedral_angles(ndx)
else:
angles = np.concatenate(
(angles, cangle.get_dihedral_angles(ndx)), axis=0)
if args.v:
print("Write angles to output file: ", args.o)
# write angles to an output file
write_angles(angles,
args.o,
cosine=args.cos,
sine=args.sin,
dt=args.dt,
begin=args.b,
end=args.e)
else:
print("Some of the input files are not existed. Input again.")
if not os.path.exists(args.n):
print("Index file is not provided. Run again.")
if not os.path.exists(args.f):
print("Xtc trajectory file is not provided. Run again.")
if not os.path.exists(args.s):
print("Reference file is not provided. Run again.")
sys.exit(1)
def iterload_cmap():
"""Load large trajectory iteratively using mdtraj.iterload function,
then generate contact map from the trajectories.
Returns
-------
"""
# for calculation time counting
startTime = datetime.now()
# argument options
args = arguments()
verbose(args.v, "Atom selecting ......")
# TODO: atom selection method required
if os.path.exists(args.s) and args.s[-4:] == ".pdb":
inp = args.s
elif os.path.exists(args.f) and args.f[-4:] == ".pdb":
inp = args.f
else:
inp = None
print("Reference pdb file is not existed. Exit now!")
sys.exit(0)
rec_index = int(args.rc[2]) - int(args.rc[1]) + 1
lig_index = int(args.lc[2]) - int(args.lc[1]) + 1
verbose(args.v, "Loading trajectory ......")
# read gromacs trajectory
trajs = gmxcli.read_xtc(args.f,
args.s,
chunk=1000,
stride=int(args.dt / args.ps))
n_frames = sum([x.n_frames for x in trajs])
verbose(args.v, "Total number of frames: %d " % n_frames)
verbose(args.v, "Start calculating contact map ......")
if args.NbyN:
contact_map = cmap_nbyn(trajs, inp, args.rc, args.lc, args.v,
args.cutoff, " ABCDEFGHIJK", args.atomtype)
else:
contact_map = cmap_general(trajs,
inp,
args.rc,
args.lc,
args.atomtype,
args.cutoff,
v=args.v,
switch=args.switch)
# subset the results
verbose(args.v, "Preparing output file ......")
contact_map = pd.DataFrame(contact_map)
contact_map.index = np.arange(contact_map.shape[0]) * args.dt
contact_map = pca.datset_subset(contact_map, args.b, args.e)
# get mean cmap data
if args.opt in ['A', 'a', 'average', 'Average']:
results = np.mean(contact_map, axis=0).reshape((rec_index, lig_index))
results = pd.DataFrame(results)
results.index = range(int(args.rc[1]), int(args.rc[2]) + 1)
results.columns = range(int(args.lc[1]), int(args.lc[2]) + 1)
else:
results = contact_map
results = pd.DataFrame(results)
results.index = np.arange(results.shape[0]) * args.dt
results.columns = [str(x) for x in np.arange(results.shape[1])]
# save results to an output file
verbose(args.v, "Writing output now ...... ")
results.to_csv(args.o,
sep=",",
header=True,
index=True,
float_format="%.3f")
print("Total Time Usage: ")
print(datetime.now() - startTime)
from mdanaly import gmxcli, pca, cmap
from dockml import index
import sys
import numpy as np
if __name__ == "__main__":
ref = sys.argv[1]
xtc = sys.argv[2]
ind = sys.argv[3]
# load xtc first
trajs = gmxcli.read_xtc(xtc, ref, 1000, 50)
# process index
ndx = index.GmxIndex(ind)
sets = ["receptor", "ligand"]
used_groups = []
atom_indices = []
for i in [0, 1]:
print("Please select a group for %s: " % sets[i])
for j, gn in enumerate(ndx.groups):
print("%d : %s" % (j, gn))
used_groups.append(ndx.groups[int(input("Your choice: "))])
rec_ndx = [int(x) - 1 for x in ndx.groupContent(used_groups[0])]
lig_ndx = [int(x) - 1 for x in ndx.groupContent(used_groups[1])]
cmaps = np.array([])