def setUp(self):
self.top = md.load(test_filenames.top_pdb).top
self.fragments_by_resSeqplus = mdcfragments.get_fragments(
self.top, method="resSeq+", verbose=False)
self.fragments_by_resSeq = mdcfragments.get_fragments(self.top,
method="resSeq",
verbose=False)
def test_works_3CAP_vs_1U19_just_runs(self):
geom_3CAP = md.load(test_filenames.pdb_3CAP)
geom_1U19 = md.load(test_filenames.pdb_1U19)
frag_3CAP = get_fragments(geom_3CAP.top,
verbose=True,
atoms=True
)[0]
frag_1U19 = get_fragments(geom_1U19.top,
verbose=False,
atoms=True
)[0]
geom_3CAP = geom_3CAP.atom_slice(frag_3CAP)
geom_1U19 = geom_1U19.atom_slice(frag_1U19)
seq_3CAP = sequence.top2seq(geom_3CAP.top)
seq_1U19 = sequence.top2seq(geom_1U19.top)
ialg = sequence.my_bioalign(seq_3CAP,
seq_1U19,
)[0]
sequence.alignment_result_to_list_of_dicts(ialg,
_np.arange(geom_3CAP.n_residues),
_np.arange(geom_1U19.n_residues),
topology_0=geom_3CAP.top,
topology_1=geom_1U19.top,
#verbose=True
)
def test_atoms(self):
for imethod in _allowed_fragment_methods:
fragments_residxs = mdcfragments.get_fragments(self.geom.top,
method=imethod)
fragments_atom_idxs = mdcfragments.get_fragments(self.geom.top,
atoms=True,
method=imethod)
for rfrag, afrag in zip(fragments_residxs, fragments_atom_idxs):
ridxs_by_atom = _np.unique(
[self.geom.top.atom(aa).residue.index for aa in afrag])
assert _np.allclose(ridxs_by_atom, rfrag)
def test_resSeq(self):
by_resSeq = mdcfragments.get_fragments(self.geom.top, method='resSeq')
assert _np.allclose(by_resSeq[0], [0, 1, 2])
assert _np.allclose(by_resSeq[1], [3, 4])
assert _np.allclose(by_resSeq[2], [5])
assert _np.allclose(by_resSeq[3], [6, 7])
def sites_to_res_pairs(
site_dicts,
top,
fragments=None,
**get_fragments_kwargs,
):
r"""Return the pairs of res_idxs needed to compute
all the contacts contained in all the input sites.
The idea is to join all needed pairs of res_idxs
in one list regardless of where they come from.
Parameters
----------
site_dicts : list of dicts
Check :obj:`x2site` for how these dicts look like
top : :obj:`~mdtraj.Topology`
fragments : list, default is None
You can pass along fragment definitions so that
it's easier to de-duplicate any AA in your input.
Otherwise, these will be created on-the-fly
by :obj:`mdciao.fragments.get_fragments`
get_fragments_kwargs :
see :obj:`fragments.get_fragments`
Returns
-------
res_idxs_pairs : 2D np.ndarray
Unique residue pairs contained in the :obj:`site_dicts`,
expressed as residue indices of :obj:`top`
[0,1] is considered != [0,1]
site_maps : list
For each site, a list with the indices of :obj:`res_idxs_pairs`
that match the site's pairs.
"""
if fragments is None:
fragments = _mdcfrg.get_fragments(top, **get_fragments_kwargs)
get_pair_lambda = {
"AAresSeq":
lambda bond: _mdcu.residue_and_atom.residues_from_descriptors(
bond, fragments, top)[0],
"residx":
lambda bond: bond
}
res_idxs_pairs = []
pair2idx = {}
site_maps = []
for ii, site in enumerate(site_dicts):
imap = []
for bond_type, bonds in site["pairs"].items():
for bond in bonds:
pair = tuple(get_pair_lambda[bond_type](bond))
if pair not in res_idxs_pairs:
res_idxs_pairs.append(pair)
pair2idx[pair] = len(res_idxs_pairs) - 1
imap.append(pair2idx[pair])
site_maps.append(imap)
#print(site_maps)
return _np.vstack(res_idxs_pairs), site_maps
def test_resSeqplus(self):
by_resSeqplus = mdcfragments.get_fragments(self.geom.top,
verbose=True,
method='resSeq+')
assert _np.allclose(by_resSeqplus[0], [0, 1, 2])
assert _np.allclose(by_resSeqplus[1], [3, 4, 5, 6, 7])
def test_bonds(self):
by_bonds = mdcfragments.get_fragments(self.geom.top, method='bonds')
assert _np.allclose(by_bonds[0], [0, 1, 2])
assert _np.allclose(by_bonds[1], [3, 4, 5])
assert _np.allclose(by_bonds[2], [6])
assert _np.allclose(by_bonds[3], [7])
def test_resSeq_file(self):
by_resSeq = mdcfragments.get_fragments(test_filenames.small_monomer,
method='resSeq')
assert _np.allclose(by_resSeq[0], [0, 1, 2])
assert _np.allclose(by_resSeq[1], [3, 4])
assert _np.allclose(by_resSeq[2], [5])
assert _np.allclose(by_resSeq[3], [6, 7])
def test_join_fragments_normal(self):
by_bonds = mdcfragments.get_fragments(self.geom.top,
join_fragments=[[1, 2]],
verbose=True,
method='bonds')
assert _np.allclose(by_bonds[0], [0, 1, 2])
assert _np.allclose(by_bonds[1], [3, 4, 5, 6])
assert _np.allclose(by_bonds[2], [7])
def test_join_fragments_special_cases(self):
# Checking if redundant fragment ids are removed from the inner list for the argument "join_fragments"
by_bonds = mdcfragments.get_fragments(self.geom.top,
join_fragments=[[1, 2, 2]],
verbose=True,
method='bonds')
assert _np.allclose(by_bonds[0], [0, 1, 2])
assert _np.allclose(by_bonds[1], [3, 4, 5, 6])
assert _np.allclose(by_bonds[2], [7])
# Checking for error from the overlapping ids in the argument "join_fragments"
with pytest.raises(AssertionError):
mdcfragments.get_fragments(self.geom.top,
join_fragments=[[1, 2], [2, 3]],
verbose=True,
method='bonds')
def test_resSeq_bonds(self):
by_both = mdcfragments.get_fragments(
self.geom.top, method='resSeq_bonds') #method is both
assert _np.allclose(by_both[0], [0, 1, 2])
assert _np.allclose(by_both[1], [3, 4])
assert _np.allclose(by_both[2], [5])
assert _np.allclose(by_both[3], [6])
assert _np.allclose(by_both[4], [7])
def setUpClass(cls):
BW_file = path.relpath(test_filenames.adrb2_human_xlsx,
test_filenames.RSCB_pdb_path)
cls.BW_local_w_pdb = nomenclature.LabelerBW(
BW_file,
ref_PDB="3SN6",
local_path=test_filenames.RSCB_pdb_path,
format="%s",
)
cls.fragments = get_fragments(cls.BW_local_w_pdb.top, verbose=False)
def test_break_fragments_special_cases_breaker_not_present(self):
# No new fragments are created if residue id is not present anywhere
by_bonds = mdcfragments.get_fragments(
self.geom.top,
fragment_breaker_fullresname=["Glu30"], # not a valid id
verbose=True,
method='bonds')
assert _np.allclose(by_bonds[0], [0, 1, 2])
assert _np.allclose(by_bonds[1], [3, 4, 5])
assert _np.allclose(by_bonds[2], [6])
assert _np.allclose(by_bonds[3], [7])
def test_with_BW(self):
fragments = mdcfragments.get_fragments(self.geom.top)
input_values = (val for val in [""])
option = test_filenames.adrb2_human_xlsx
with mock.patch('builtins.input', lambda *x: next(input_values)):
residx2conlab, lblr = cli._parse_consensus_option(option, "BW",
self.geom.top,
fragments,
return_Labeler=True,
try_web_lookup=False)
self.assertIsInstance(lblr, LabelerBW)
self.assertIsInstance(residx2conlab,list)
def test_break_fragments_special_cases_already_breaker_passed_as_string(
self):
# Also works if input is a string instead of an iterable of strings
by_bonds = mdcfragments.get_fragments(
self.geom.top,
fragment_breaker_fullresname=
"GLU30", # GLU30 is already a fragment breaker
verbose=True,
method='bonds')
assert _np.allclose(by_bonds[0], [0, 1, 2])
assert _np.allclose(by_bonds[1], [3, 4, 5])
assert _np.allclose(by_bonds[2], [6])
assert _np.allclose(by_bonds[3], [7])
def test_break_fragments_special_cases_already_breaker(self):
# No new fragments are created if an existing fragment breaker is passed
by_bonds = mdcfragments.get_fragments(
self.geom.top,
fragment_breaker_fullresname=[
"GLU30"
], # GLU30 is already a fragment breaker
verbose=True,
method='bonds')
assert _np.allclose(by_bonds[0], [0, 1, 2])
assert _np.allclose(by_bonds[1], [3, 4, 5])
assert _np.allclose(by_bonds[2], [6])
assert _np.allclose(by_bonds[3], [7])
def test_with_BW_already_instantiated(self):
fragments = mdcfragments.get_fragments(self.geom.top)
BW = LabelerBW(test_filenames.adrb2_human_xlsx)
input_values = (val for val in [""])
with mock.patch('builtins.input', lambda *x: next(input_values)):
residx2conlab, lblr = cli._parse_consensus_option(BW, "BW",
self.geom.top,
fragments,
return_Labeler=True)
self.assertIsInstance(lblr, LabelerBW)
self.assertEqual(lblr,BW)
self.assertIsInstance(residx2conlab,list)
def test_break_fragments(self):
# Checking if the fragments are breaking correctly for the argument "fragment_breaker_fullresname"
by_bonds = mdcfragments.get_fragments(
self.geom.top,
fragment_breaker_fullresname=["VAL31", "GLU27"
], # two fragment breakers passed
verbose=True,
method='bonds')
assert _np.allclose(by_bonds[0], [0])
assert _np.allclose(by_bonds[1], [1, 2])
assert _np.allclose(by_bonds[2], [3])
assert _np.allclose(by_bonds[3], [4, 5])
assert _np.allclose(by_bonds[4], [6])
assert _np.allclose(by_bonds[5], [7])
def test_chains(self):
by_chains = mdcfragments.get_fragments(self.geom.top, method="chains")
assert _np.allclose(by_chains[0], [0, 1, 2])
assert _np.allclose(by_chains[1], [3, 4, 5])
assert _np.allclose(by_chains[2], [6, 7]), by_chains
def setUp(self):
self.top = md.load(test_filenames.small_monomer).top
self.fragments = _mdcfrg.get_fragments(self.top, method="resSeq+",
verbose=False)
def setUp(self):
self.geom2frags = md.load(test_filenames.small_dimer)
self.by_bonds_geom2frags = _mdcfrg.get_fragments(self.geom2frags.top,
verbose=True,
method='bonds')
def test_Nonestr(self):
nonefrags = mdcfragments.get_fragments(self.geom.top, method="None")
_np.testing.assert_array_equal(nonefrags[0],
_np.arange(self.geom.top.n_residues))
self.assertEqual(len(nonefrags), 1)
def setUp(self):
self.geom = md.load(test_filenames.top_pdb)
self.fragments = mdcfragments.get_fragments(self.geom.top,
verbose=False)
def test_dont_know_method(self):
with pytest.raises(AssertionError):
mdcfragments.get_fragments(self.geom.top, method='xyz')
def test_ions_and_water(self):
geom = md.load(test_filenames.ions_and_water)
frags = mdcfragments.get_fragments(geom.top, verbose=False)
assert len(frags) == 3
assert len(frags[1]) == 5
assert len(frags[2]) == 10
def test_molecule_resSeqplus_raises_not_imp(self):
with pytest.raises(NotImplementedError):
mdcfragments.get_fragments(self.geom.top,
method="molecules_resSeq+")
def test_completeness(self):
for imethod in _allowed_fragment_methods:
frags = mdcfragments.get_fragments(self.geom.top, method=imethod)
assert _np.allclose(_np.hstack(frags),
_np.arange(self.geom.top.n_residues))
