def example_datasets(
draw,
min_size: int = 60,
max_size: int = 130,
directory: Path = Path("."),
) -> MultiCrystalDataset:
dataset_paths_dict = parse_pandda_input(directory)
paths_strategy = hypothesis.strategies.sampled_from(
[x for x in dataset_paths_dict.keys()])
dataset_paths_subset = draw(
hypothesis.strategies.lists(
paths_strategy,
unique=True,
min_size=min_size,
max_size=max_size,
))
datasets = {
dtag: Dataset(
reflections_from_mtz(MTZFile(
dataset_paths_dict[dtag]["mtz_path"])),
structure_biopandas_from_pdb(
PDBFile(dataset_paths_dict[dtag]["pdb_path"])))
for dtag in dataset_paths_subset
}
return MultiCrystalDataset(datasets)
def normalise_structure(reference_pdb_path, dtag, output_path):
f = PDBFile(reference_pdb_path)
structure = structure_biopython_from_pdb(f)
box_origin = np.min(np.vstack(
[atom.coord for atom in structure.structure.get_atoms()]),
axis=0)
print("\tBox origin is: {}".format(box_origin))
translated_structure = translate_structure(structure, np.eye(3),
-box_origin)
box_origin = np.min(np.vstack(
[atom.coord for atom in translated_structure.structure.get_atoms()]),
axis=0)
print(box_origin)
translated_structure.output(output_path / "{}_normalised.pdb".format(dtag))
def align_map_to_reference(
dtag,
reference_dtag,
dataset_path,
reference_pdb_path,
output_path,
min_res,
structure_factors="FWT,PHWT",
):
# Load structures
f_ref = PDBFile(reference_pdb_path)
reference_structure = structure_biopython_from_pdb(f_ref)
f_moving = PDBFile(dataset_path["pdb_path"])
moving_structure = structure_biopython_from_pdb(f_moving)
# Load xmap
xmap = mdc3.types.real_space.xmap_from_path(
dataset_path["mtz_path"],
structure_factors,
)
# Get box limits from reference structure
box_limits = np.max(
np.vstack([
atom.coord for atom in reference_structure.structure.get_atoms()
]),
axis=0,
)
print(box_limits)
# Align and Get RTop to moving protein frame from alignment
alignment_moving_to_ref = mdc3.functions.alignment.align(
reference_structure.structure,
moving_structure.structure,
)
alignment_ref_to_moving = mdc3.functions.alignment.align(
moving_structure.structure,
reference_structure.structure,
)
rotation = alignment_moving_to_ref.rotran[0]
translation = alignment_moving_to_ref.rotran[1]
alignment_moving_to_ref.apply(moving_structure.structure)
print("translation: orthogonal to grid")
print(translation)
print("rotation")
print(rotation)
# Interpolate NX map in moving protein frame
grid_params = [int(x) + 5 for x in box_limits]
nxmap = mdc3.types.real_space.interpolate_uniform_grid(
xmap,
translation,
np.transpose(rotation),
grid_params=grid_params,
)
nxmap_data = nxmap.export_numpy()
origin_nxmap = clipper_python.NXmap_float(
clipper_python.Grid(
grid_params[0],
grid_params[1],
grid_params[2],
),
clipper_python.RTop_orth(
clipper_python.Mat33_double(np.eye(3)),
clipper_python.Vec3_double(0, 0, 0),
))
origin_nxmap.import_numpy(
clipper_python.Coord_grid(0, 0, 0),
nxmap_data,
)
# Output to ccp4
# cell = xmap.xmap.cell
cell = clipper_python.Cell(
clipper_python.Cell_descr(
grid_params[0],
grid_params[1],
grid_params[2],
np.pi / 2,
np.pi / 2,
np.pi / 2,
))
mdc3.types.real_space.output_nxmap(
origin_nxmap,
output_path / "{}_origin.ccp4".format(dtag),
cell,
)
mdc3.types.real_space.output_nxmap(
nxmap,
output_path / "{}.ccp4".format(dtag),
cell,
)
# Output aligned pdb
moving_structure.output(output_path / "{}_aligned.pdb".format(dtag))
def mtz_to_ccp4(
self,
dtag,
reference_pdb_path,
dataset_path,
output_path,
min_res,
structure_factors="FWT,PHWT",
):
# Load structures
f_ref = PDBFile(reference_pdb_path)
reference_structure = structure_biopython_from_pdb(f_ref)
# Load xmap
# xmap = mdc3.types.real_space.xmap_from_path(dataset_path["mtz_path"],
# structure_factors,
# )
# Get box limits from reference structure
box_limits_max = np.max(
np.vstack([
atom.coord
for atom in reference_structure.structure.get_atoms()
]),
axis=0,
)
box_limits_min = np.min(
np.vstack([
atom.coord
for atom in reference_structure.structure.get_atoms()
]),
axis=0,
)
# Interpolate NX map in moving protein frame
grid_params = [int(x) + 4 for x in (box_limits_max - box_limits_min)]
# nxmap = mdc3.types.real_space.interpolate_uniform_grid(xmap,
# box_limits_min - np.array([2, 2, 2]),
# np.eye(3),
# grid_params=grid_params,
# )
#
# # Output to ccp4
# cell = clipper_python.Cell(clipper_python.Cell_descr(grid_params[0],
# grid_params[1],
# grid_params[2],
# np.pi / 2,
# np.pi / 2,
# np.pi / 2,
# )
# )
#
# mdc3.types.real_space.output_nxmap(nxmap,
# output_path / "{}.ccp4".format(dtag),
# cell,
# )
mtz = gemmi.read_mtz_file(str(dataset_path["mtz_path"]))
all_data = np.array(mtz, copy=False)
mtz.set_data(all_data[mtz.make_d_array() >= 3.0])
grid = mtz.transform_f_phi_to_map(
"FWT",
"PHWT",
sample_rate=3,
)
mp = gemmi.Ccp4Map()
mp.grid = grid
mp.update_ccp4_header(2, True)
mp.write_ccp4_map(str(output_path / "{}.ccp4".format(dtag)))
from hypothesis import given, settings
from hypothesis.strategies import just
from pathlib import Path
from mdc3.types.files import (PDBFile,
MTZFile,
)
from mdc3.types.structures import structure_biopandas_from_pdb
from mdc3.types.reflections import (Reflections,
reflections_from_mtz,
new_reflections_from_reflections_at_res,
)
@given(just(PDBFile(Path("../../data/dimple.pdb").resolve())))
def test_structure_biopandas_from_pdb(pdb_file: PDBFile) -> None:
structure_biopandas_from_pdb(pdb_file)
@given(just(MTZFile(Path("../../data/dimple.mtz").resolve())))
def test_reflections_from_mtz(mtz_file: MTZFile) -> None:
reflections_from_mtz(mtz_file)
@settings(deadline=1000)
@given(just(reflections_from_mtz(MTZFile(Path("../../data/dimple.mtz").resolve()))),
just(2),
)
def test_new_reflections_from_reflections_at_res(reflections: Reflections,
resolution: float,