def read(self, f_handle, f_id="casp2"):
"""Read a distance prediction file
Parameters
----------
f_handle
Open file handle [read permissions]
f_id : str, optional
Unique contact file identifier
Returns
-------
:obj:`~conkit.core.distancefile.DistanceFile`
"""
hierarchy = DistanceFile(f_id)
hierarchy.original_file_format = "CASPRR_MODE_2"
_map = Distogram("distogram_1")
hierarchy.add(_map)
for line in f_handle.readlines():
line = line.lstrip().rstrip().split()
if not line or len(line) != 13 or not line[0].isdigit() or not line[1].isdigit():
continue
res1_seq = int(line[0])
res2_seq = int(line[1])
raw_score = float(line[2])
distance_scores = tuple([float(p) for p in line[3:]])
_distance = Distance(res1_seq, res2_seq, distance_scores, DISTANCE_BINS, raw_score=raw_score)
_map.add(_distance)
return hierarchy
def test_original_file_format(self):
distance_file = DistanceFile("test")
distance_file.original_file_format = "pdb"
distogram = Distogram("test")
distance_file.add(distogram)
self.assertTrue(distogram in distance_file.child_list)
self.assertEqual("pdb", distogram.original_file_format)
def read(self, f_handle, f_id="rosettanpz"):
"""Read a distance prediction file
Parameters
----------
f_handle
Open file handle [read permissions]
f_id : str, optional
Unique contact file identifier
Returns
-------
:obj:`~conkit.core.distancefile.DistanceFile`
"""
hierarchy = DistanceFile(f_id)
hierarchy.original_file_format = "ROSETTA_NPZ"
_map = Distogram("distogram_1")
hierarchy.add(_map)
prediction = np.load(f_handle, allow_pickle=True)
probs = prediction['dist']
# Bin #0 corresponds with d>20A & bins #1 ~ #36 correspond with 2A<d<20A in increments of 0.5A
probs = probs[:, :, [x for x in range(1, 37)] + [0]]
L = probs.shape[0]
for i in range(L):
for j in range(i, L):
_distance = Distance(i + 1, j + 1, tuple(probs[i, j, :].tolist()), DISTANCE_BINS)
_map.add(_distance)
return hierarchy
def test_write_1(self):
expected_output = """PFRMAT RR
RMODE 2
1 6 0.199696 0.043889 0.085795 0.070011 0.071518 0.054028 0.213284 0.069087 0.097959 0.090083 0.204345
1 7 0.233644 0.049411 0.075135 0.109098 0.150810 0.096584 0.092398 0.096662 0.093350 0.123176 0.113375
1 8 0.246451 0.106886 0.039024 0.100540 0.082028 0.108344 0.078788 0.105980 0.130109 0.113708 0.134592
1 9 0.267139 0.072002 0.083053 0.112084 0.124356 0.128044 0.097491 0.132106 0.047198 0.110915 0.092751
1 10 0.351914 0.081445 0.069721 0.200748 0.099755 0.090368 0.117449 0.127677 0.050879 0.101965 0.059993
2 7 0.228459 0.085973 0.091366 0.051120 0.085890 0.070657 0.119253 0.082744 0.180051 0.097734 0.135213
2 8 0.256177 0.081094 0.077748 0.097335 0.060811 0.138077 0.130496 0.106911 0.101101 0.121346 0.085081
2 9 0.216631 0.046454 0.053018 0.117160 0.196036 0.144154 0.125199 0.090720 0.052621 0.098583 0.076055
2 10 0.284653 0.087567 0.125308 0.071778 0.071988 0.095966 0.099270 0.174715 0.109563 0.062611 0.101233
3 8 0.345583 0.117500 0.110134 0.117950 0.085312 0.098812 0.072826 0.079326 0.196758 0.059058 0.062325
3 9 0.203586 0.036574 0.050725 0.116287 0.174339 0.070881 0.116388 0.083683 0.060738 0.160257 0.130128
3 10 0.293849 0.059364 0.135117 0.099368 0.113124 0.135930 0.066876 0.075962 0.114771 0.127034 0.072454
4 9 0.234649 0.077170 0.048841 0.108638 0.107559 0.119732 0.116349 0.077063 0.111788 0.119497 0.113362
4 10 0.322930 0.090789 0.133412 0.098729 0.099123 0.084633 0.107534 0.137072 0.096560 0.042234 0.109913
5 10 0.279782 0.054314 0.114427 0.111042 0.069073 0.083048 0.105829 0.073806 0.119769 0.088666 0.180028"""
distancefile = DistanceFile("test")
distancefile.original_file_format = 'ALPHAFOLD2'
distogram = Distogram("1")
distancefile.add(distogram)
list_res1 = [1, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 4, 4, 5]
list_res2 = [6, 7, 8, 9, 10, 7, 8, 9, 10, 8, 9, 10, 9, 10, 10]
bin_edges = (2.3125, 2.625, 2.9375, 3.25, 3.5625, 3.875, 4.1875, 4.5,
4.8125, 5.125, 5.4375, 5.75, 6.0625, 6.375, 6.6875,
6.9999995, 7.3125, 7.625, 7.9375, 8.25, 8.5625, 8.875,
9.1875, 9.5, 9.812499, 10.124999, 10.4375, 10.75, 11.0625,
11.375, 11.687499, 12., 12.3125, 12.625, 12.9375, 13.25,
13.5625, 13.874999, 14.187501, 14.499999, 14.812499,
15.124999, 15.437499, 15.75, 16.0625, 16.375, 16.687502,
16.999998, 17.312498, 17.624998, 17.937498, 18.25,
18.5625, 18.875, 19.1875, 19.5, 19.8125, 20.125,
20.437498, 20.75, 21.062498, 21.374998, 21.6875)
distance_bins = [(0, bin_edges[0])]
distance_bins += [(bin_edges[idx], bin_edges[idx + 1])
for idx in range(len(bin_edges) - 1)]
distance_bins.append((bin_edges[-1], np.inf))
distance_bins = tuple(distance_bins)
np.random.seed(41)
for res_1, res_2 in zip(list_res1, list_res2):
distance_scores = np.random.dirichlet(np.ones(64)).tolist()
distance = Distance(res_1, res_2, distance_scores, distance_bins)
distogram.add(distance)
f_name = self.tempfile()
with open(f_name, "w") as f_out:
CaspMode2Parser().write(f_out, distogram)
with open(f_name, "r") as f_in:
output = f_in.read().splitlines()
self.assertListEqual(expected_output.split('\n'), output)
def test_write_1(self):
expected_output = """#REMARK MapPred 1.1
#REMARK idx_i, idx_j, distance distribution of 34 bins
#REMARK 34 bins consist of 32 normal bins (4-20A with a step of 0.5A) and two boundary bins ( [0,4) and [20, inf) ), as follows: [0,4,4.5,5,5.5,6,6.5,7,7.5,8,8.5,9,9.5,10,10.5,11,11.5,12,12.5,13,13.5,14,14.5,15,15.5,16,16.5,17,17.5,18,18.5,19,19.5,20,inf]
5 10 0.013746 0.002245 0.053742 0.002115 0.005889 0.044058 0.010081 0.052535 0.118677 0.025818 0.019215 0.015831 0.009808 0.018148 0.031220 0.003428 0.058081 0.017978 0.065069 0.024163 0.044585 0.062025 0.026062 0.023824 0.012573 0.027729 0.022212 0.041685 0.005015 0.064340 0.004133 0.006420 0.018552 0.048998
1 35 0.187103 0.008180 0.021642 0.051089 0.038619 0.006100 0.010553 0.031697 0.010831 0.015310 0.006949 0.008237 0.043400 0.051436 0.003820 0.008148 0.018467 0.057307 0.022873 0.029184 0.008235 0.008025 0.004214 0.027027 0.070948 0.028355 0.049284 0.060124 0.041885 0.043900 0.000681 0.006836 0.007679 0.011862
43 85 0.024968 0.014838 0.021987 0.031265 0.019144 0.033038 0.018177 0.008716 0.017331 0.046459 0.051147 0.043912 0.004041 0.007990 0.027690 0.073997 0.001269 0.008161 0.067709 0.055700 0.028615 0.091884 0.021842 0.025949 0.025295 0.006136 0.031655 0.028990 0.082802 0.005069 0.002322 0.015611 0.039637 0.016654
85 43 0.015871 0.013765 0.006593 0.014670 0.029273 0.042705 0.058513 0.014858 0.050493 0.014216 0.010146 0.037020 0.018679 0.003142 0.031215 0.011736 0.008920 0.007325 0.144325 0.003512 0.018591 0.005043 0.001607 0.043659 0.068744 0.052532 0.050643 0.039295 0.003413 0.035119 0.102032 0.004150 0.005737 0.032456
50 50 0.000490 0.027392 0.001090 0.009625 0.011421 0.002011 0.015100 0.018622 0.008785 0.114531 0.044962 0.019562 0.022973 0.008111 0.042691 0.061367 0.001060 0.032753 0.073944 0.006790 0.002509 0.073759 0.025060 0.031361 0.039123 0.043318 0.032752 0.004280 0.044655 0.000556 0.000111 0.095043 0.028036 0.056157
18 50 0.002704 0.015000 0.024442 0.105520 0.014259 0.027628 0.002832 0.035063 0.038354 0.055931 0.039683 0.035546 0.004621 0.019932 0.012316 0.087781 0.006637 0.043857 0.008459 0.053482 0.016937 0.083507 0.031733 0.000793 0.004304 0.066937 0.009968 0.006859 0.038950 0.064003 0.003185 0.008042 0.007331 0.023401"""
distancefile = DistanceFile("test")
distancefile.original_file_format = 'MAPPRED'
distogram = Distogram("1")
distancefile.add(distogram)
list_res1 = [5, 1, 43, 85, 50, 18]
list_res2 = [10, 35, 85, 43, 50, 50]
distance_bins = ((0, 4), (4, 4.5), (4.5, 5), (5, 5.5), (5.5, 6),
(6, 6.5), (6.5, 7), (7, 7.5), (7.5, 8), (8, 8.5),
(8.5, 9), (9, 9.5), (9.5, 10), (10, 10.5), (10.5, 11),
(11, 11.5), (11.5, 12), (12, 12.5), (12.5, 13),
(13, 13.5), (13.5, 14), (14, 14.5), (14.5, 15),
(15, 15.5), (15.5, 16), (16, 16.5), (16.5, 17),
(17, 17.5), (17.5, 18), (18, 18.5), (18.5, 19),
(19, 19.5), (19.5, 20), (20, np.inf))
np.random.seed(41)
for res_1, res_2 in zip(list_res1, list_res2):
distance_scores = np.random.dirichlet(np.ones(34)).tolist()
distance = Distance(res_1, res_2, distance_scores, distance_bins)
distogram.add(distance)
f_name = self.tempfile()
with open(f_name, "w") as f_out:
MapPredParser().write(f_out, distogram)
with open(f_name, "r") as f_in:
output = f_in.read().splitlines()
self.assertListEqual(expected_output.split("\n"), output)
def read(self, f_handle, f_id="alphafold2"):
"""Read a distance prediction file
Parameters
----------
f_handle
Open file handle [read permissions]
f_id : str, optional
Unique contact file identifier
Returns
-------
:obj:`~conkit.core.distancefile.DistanceFile`
"""
hierarchy = DistanceFile(f_id)
hierarchy.original_file_format = "alphafold2"
_map = Distogram("distogram_1")
hierarchy.add(_map)
prediction = np.load(f_handle, allow_pickle=True)
predicted_distogram = prediction['distogram']
probs = softmax(predicted_distogram['logits'], axis=-1)
bin_edges = predicted_distogram['bin_edges']
distance_bins = [(0, bin_edges[0])]
distance_bins += [(bin_edges[idx], bin_edges[idx + 1]) for idx in range(len(bin_edges) - 1)]
distance_bins.append((bin_edges[-1], np.inf))
distance_bins = tuple(distance_bins)
L = probs.shape[0]
for i in range(L):
for j in range(i, L):
_distance = Distance(i + 1, j + 1, tuple(probs[i, j, :].tolist()), distance_bins)
_map.add(_distance)
return hierarchy
def DistanceFile(*args, **kwargs):
""":obj:`Contact <conkit.core.distancefile.DistanceFile>` instance"""
from conkit.core.distancefile import DistanceFile
return DistanceFile(*args, **kwargs)
def _read(self, structure, f_id, distance_cutoff, atom_type):
"""Read a contact file
Parameters
----------
structure
A :obj:`~Bio.PDB.Structure.Structure>` instance
f_id : str
Unique contact file identifier
distance_cutoff : int
Distance cutoff for which to determine contacts
atom_type : str
Atom type between which distances are calculated
Returns
-------
:obj:`~conkit.core.distancefile.DistanceFile~`
"""
hierarchies = []
distance_bound = (0.0, float(distance_cutoff))
for model in structure:
hierarchy = DistanceFile(f_id + "_" + str(model.id))
hierarchy.original_file_format = "PDB"
chains = list(chain for chain in model)
for chain in chains:
self._remove_hetatm(chain)
self._remove_atom(chain, atom_type)
for chain1, chain2 in itertools.product(chains, chains):
if chain1.id == chain2.id: # intra
distogram = Distogram(chain1.id)
else: # inter
distogram = Distogram(chain1.id + chain2.id)
for (atom1, atom2,
distance) in self._chain_contacts(chain1, chain2):
if distance < distance_cutoff:
score = round(1.0 - (distance / 100), 6)
else:
score = 0
dist = Distance(atom1.resseq, atom2.resseq, (1, ),
((distance, distance), ), score,
distance_bound)
dist.res1_altseq = atom1.resseq_alt
dist.res2_altseq = atom2.resseq_alt
dist.res1 = atom1.resname
dist.res2 = atom2.resname
dist.res1_chain = atom1.reschain
dist.res2_chain = atom2.reschain
if distance_cutoff == 0 or distance < distance_cutoff:
dist.true_positive = True
distogram.add(dist)
if distogram.empty:
del distogram
else:
if len(distogram.id) == 1:
distogram.sequence = self._build_sequence(chain1)
assert len(distogram.sequence.seq) == len(chain1)
else:
distogram.sequence = self._build_sequence(
chain1) + self._build_sequence(chain2)
assert len(distogram.sequence.seq
) == len(chain1) + len(chain2)
hierarchy.add(distogram)
hierarchy.method = "Distogram extracted from PDB " + str(model.id)
hierarchy.remark = [
"The model id is the chain identifier, i.e XY equates to chain X and chain Y.",
"Residue numbers in column 1 are chain X, and numbers in column 2 are chain Y.",
]
hierarchies.append(hierarchy)
if len(hierarchies) > 1:
msg = "Super-level to contact file not yet implemented. " "Parser returns hierarchy for top model only!"
warnings.warn(msg, FutureWarning)
return hierarchies[0]