| 1 |
""" |
| 2 |
PDB structure parsers, format builders and database providers. |
| 3 |
|
| 4 |
The most basic usage is: |
| 5 |
|
| 6 |
>>> parser = StructureParser('structure.pdb') |
| 7 |
>>> parser.parse_structure() |
| 8 |
<Structure> # a Structure object (model) |
| 9 |
|
| 10 |
or if this is an NMR ensemble: |
| 11 |
|
| 12 |
>>> parser.parse_models() |
| 13 |
<Ensemble> # an Ensemble object (collection of alternative Structure-s) |
| 14 |
|
| 15 |
This module introduces a family of PDB file parsers. The common interface of all |
| 16 |
parsers is defined in L{AbstractStructureParser}. This class has several |
| 17 |
implementations: |
| 18 |
|
| 19 |
- L{RegularStructureParser} - handles normal PDB files with SEQRES fields |
| 20 |
- L{LegacyStructureParser} - reads structures from legacy or malformed PDB files, |
| 21 |
which are lacking SEQRES records (initializes all residues from the ATOMs instead) |
| 22 |
- L{PDBHeaderParser} - reads only the headers of the PDB files and produces structures |
| 23 |
without coordinates. Useful for reading metadata (e.g. accession numbers or just |
| 24 |
plain SEQRES sequences) with minimum overhead |
| 25 |
|
| 26 |
Unless you have a special reason, you should use the L{StructureParser} factory, |
| 27 |
which returns a proper L{AbstractStructureParser} implementation, depending on the |
| 28 |
input PDB file. If the input file looks like a regular PDB file, the factory |
| 29 |
returns a L{RegularStructureParser}, otherwise it instantiates L{LegacyStructureParser}. |
| 30 |
L{StructureParser} is in fact an alias for L{AbstractStructureParser.create_parser}. |
| 31 |
|
| 32 |
Writing your own, customized PDB parser is easy. Suppose that you are trying to |
| 33 |
parse a PDB-like file which misuses the charge column to store custom info. This |
| 34 |
will certainly crash L{RegularStructureParser} (for good), but you can create your |
| 35 |
own parser as a workaround. All you need to to is to override the virtual |
| 36 |
C{_read_charge} hook method:: |
| 37 |
|
| 38 |
class CustomParser(RegularStructureParser): |
| 39 |
|
| 40 |
def _read_charge(self, line): |
| 41 |
try: |
| 42 |
return super(CustomParser, self)._read_charge(line) |
| 43 |
except StructureFormatError: |
| 44 |
return None |
| 45 |
|
| 46 |
Another important abstraction in this module is L{StructureProvider}. It has several |
| 47 |
implementations which can be used to retrieve PDB L{Structure}s from various sources: |
| 48 |
file system directories, remote URLs, etc. You can easily create your own provider as |
| 49 |
well. See L{StructureProvider} for details. |
| 50 |
|
| 51 |
Finally, this module gives you some L{FileBuilder}s, used for text serialization |
| 52 |
of L{Structure}s and L{Ensemble}s: |
| 53 |
|
| 54 |
>>> builder = PDBFileBuilder(stream) |
| 55 |
>>> builder.add_header(structure) |
| 56 |
>>> builder.add_structure(structure) |
| 57 |
|
| 58 |
where stream is any Python stream, e.g. an open file or sys.stdout. |
| 59 |
|
| 60 |
See L{Ensemble} and L{Structure} from L{csb.bio.structure} for details on these |
| 61 |
objects. |
| 62 |
""" |
| 63 |
|
| 64 |
import re |
| 65 |
import os |
| 66 |
import numpy |
| 67 |
import datetime |
| 68 |
import multiprocessing |
| 69 |
|
| 70 |
import csb.bio.structure |
| 71 |
import csb.bio.sequence |
| 72 |
import csb.bio.sequence.alignment as alignment |
| 73 |
import csb.core |
| 74 |
import csb.io |
| 75 |
|
| 76 |
from abc import ABCMeta, abstractmethod |
| 77 |
from csb.bio.sequence import SequenceTypes, SequenceAlphabets |
| 78 |
from csb.bio.structure import ChemElements, SecStructures |
| 79 |
|
| 80 |
|
| 81 |
PDB_AMINOACIDS = { |
| 82 |
'PAQ': 'TYR', 'AGM': 'ARG', 'ILE': 'ILE', 'PR3': 'CYS', 'GLN': 'GLN', |
| 83 |
'DVA': 'VAL', 'CCS': 'CYS', 'ACL': 'ARG', 'GLX': 'GLX', 'GLY': 'GLY', |
| 84 |
'GLZ': 'GLY', 'DTH': 'THR', 'OAS': 'SER', 'C6C': 'CYS', 'NEM': 'HIS', |
| 85 |
'DLY': 'LYS', 'MIS': 'SER', 'SMC': 'CYS', 'GLU': 'GLU', 'NEP': 'HIS', |
| 86 |
'BCS': 'CYS', 'ASQ': 'ASP', 'ASP': 'ASP', 'SCY': 'CYS', 'SER': 'SER', |
| 87 |
'LYS': 'LYS', 'SAC': 'SER', 'PRO': 'PRO', 'ASX': 'ASX', 'DGN': 'GLN', |
| 88 |
'DGL': 'GLU', 'MHS': 'HIS', 'ASB': 'ASP', 'ASA': 'ASP', 'NLE': 'LEU', |
| 89 |
'DCY': 'CYS', 'ASK': 'ASP', 'GGL': 'GLU', 'STY': 'TYR', 'SEL': 'SER', |
| 90 |
'CGU': 'GLU', 'ASN': 'ASN', 'ASL': 'ASP', 'LTR': 'TRP', 'DAR': 'ARG', |
| 91 |
'VAL': 'VAL', 'CHG': 'ALA', 'TPO': 'THR', 'CLE': 'LEU', 'GMA': 'GLU', |
| 92 |
'HAC': 'ALA', 'AYA': 'ALA', 'THR': 'THR', 'TIH': 'ALA', 'SVA': 'SER', |
| 93 |
'MVA': 'VAL', 'SAR': 'GLY', 'LYZ': 'LYS', 'BNN': 'ALA', '5HP': 'GLU', |
| 94 |
'IIL': 'ILE', 'SHR': 'LYS', 'HAR': 'ARG', 'FME': 'MET', 'ALO': 'THR', |
| 95 |
'PHI': 'PHE', 'ALM': 'ALA', 'PHL': 'PHE', 'MEN': 'ASN', 'TPQ': 'ALA', |
| 96 |
'GSC': 'GLY', 'PHE': 'PHE', 'ALA': 'ALA', 'MAA': 'ALA', 'MET': 'MET', |
| 97 |
'UNK': 'UNK', 'LEU': 'LEU', 'ALY': 'LYS', 'SET': 'SER', 'GL3': 'GLY', |
| 98 |
'TRG': 'LYS', 'CXM': 'MET', 'TYR': 'TYR', 'SCS': 'CYS', 'DIL': 'ILE', |
| 99 |
'TYQ': 'TYR', '3AH': 'HIS', 'DPR': 'PRO', 'PRR': 'ALA', 'CME': 'CYS', |
| 100 |
'IYR': 'TYR', 'CY1': 'CYS', 'TYY': 'TYR', 'HYP': 'PRO', 'DTY': 'TYR', |
| 101 |
'2AS': 'ASP', 'DTR': 'TRP', 'FLA': 'ALA', 'DPN': 'PHE', 'DIV': 'VAL', |
| 102 |
'PCA': 'GLU', 'MSE': 'MET', 'MSA': 'GLY', 'AIB': 'ALA', 'CYS': 'CYS', |
| 103 |
'NLP': 'LEU', 'CYQ': 'CYS', 'HIS': 'HIS', 'DLE': 'LEU', 'CEA': 'CYS', |
| 104 |
'DAL': 'ALA', 'LLP': 'LYS', 'DAH': 'PHE', 'HMR': 'ARG', 'TRO': 'TRP', |
| 105 |
'HIC': 'HIS', 'CYG': 'CYS', 'BMT': 'THR', 'DAS': 'ASP', 'TYB': 'TYR', |
| 106 |
'BUC': 'CYS', 'PEC': 'CYS', 'BUG': 'LEU', 'CYM': 'CYS', 'NLN': 'LEU', |
| 107 |
'CY3': 'CYS', 'HIP': 'HIS', 'CSO': 'CYS', 'TPL': 'TRP', 'LYM': 'LYS', |
| 108 |
'DHI': 'HIS', 'MLE': 'LEU', 'CSD': 'ALA', 'HPQ': 'PHE', 'MPQ': 'GLY', |
| 109 |
'LLY': 'LYS', 'DHA': 'ALA', 'DSN': 'SER', 'SOC': 'CYS', 'CSX': 'CYS', |
| 110 |
'OMT': 'MET', 'DSP': 'ASP', 'PTR': 'TYR', 'TRP': 'TRP', 'CSW': 'CYS', |
| 111 |
'EFC': 'CYS', 'CSP': 'CYS', 'CSS': 'CYS', 'SCH': 'CYS', 'OCS': 'CYS', |
| 112 |
'NMC': 'GLY', 'SEP': 'SER', 'BHD': 'ASP', 'KCX': 'LYS', 'SHC': 'CYS', |
| 113 |
'C5C': 'CYS', 'HTR': 'TRP', 'ARG': 'ARG', 'TYS': 'TYR', 'ARM': 'ARG', |
| 114 |
'DNP': 'ALA' |
| 115 |
} |
| 116 |
""" |
| 117 |
Dictionary of non-standard amino acids, which could be found in PDB. |
| 118 |
""" |
| 119 |
|
| 120 |
|
| 121 |
PDB_NUCLEOTIDES = { |
| 122 |
'DA': 'Adenine', 'DG': 'Guanine', 'DC': 'Cytosine', 'DT': 'Thymine', |
| 123 |
'A': 'Adenine', 'G': 'Guanine', 'C': 'Cytosine', 'T': 'Thymine', |
| 124 |
'U': 'Uracil', 'DOC': 'Cytosine', 'R': 'Purine', 'Y': 'Pyrimidine', |
| 125 |
'K': 'Ketone', ' M': 'Amino', 'S': 'Strong', 'W': 'Weak', |
| 126 |
'B': 'NotA', 'D' : 'NotC', 'H': 'NotG', 'V': 'NotT', |
| 127 |
'N': 'Any', 'X' : 'Masked' |
| 128 |
} |
| 129 |
""" |
| 130 |
Dictionary of non-standard nucleotides, which could be found in PDB. |
| 131 |
""" |
| 132 |
|
| 133 |
class PDBParseError(ValueError): |
| 134 |
pass |
| 135 |
|
| 136 |
class HeaderFormatError(PDBParseError): |
| 137 |
pass |
| 138 |
|
| 139 |
class SecStructureFormatError(PDBParseError): |
| 140 |
pass |
| 141 |
|
| 142 |
class StructureFormatError(PDBParseError): |
| 143 |
pass |
| 144 |
|
| 145 |
class UnknownPDBResidueError(PDBParseError): |
| 146 |
pass |
| 147 |
|
| 148 |
class StructureNotFoundError(KeyError): |
| 149 |
pass |
| 150 |
|
| 151 |
class InvalidEntryIDError(StructureFormatError): |
| 152 |
pass |
| 153 |
|
| 154 |
class ResidueMappingError(StructureFormatError): |
| 155 |
pass |
| 156 |
|
| 157 |
|
| 158 |
class EntryID(object): |
| 159 |
""" |
| 160 |
Represents a PDB Chain identifier. Implementing classes must define |
| 161 |
how the original ID is split into accession number and chain ID. |
| 162 |
|
| 163 |
@param id: identifier |
| 164 |
@type id: str |
| 165 |
""" |
| 166 |
__metaclass__ = ABCMeta |
| 167 |
|
| 168 |
def __init__(self, id): |
| 169 |
|
| 170 |
self._accession = '' |
| 171 |
self._chain = '' |
| 172 |
|
| 173 |
id = id.strip() |
| 174 |
self._accession, self._chain = self.parse(id) |
| 175 |
|
| 176 |
@staticmethod |
| 177 |
def create(id): |
| 178 |
""" |
| 179 |
Guess the format of C{id} and parse it. |
| 180 |
|
| 181 |
@return: a new PDB ID of the appropriate type |
| 182 |
@rtype: L{EntryID} |
| 183 |
""" |
| 184 |
|
| 185 |
if len(id) in (4, 5): |
| 186 |
return StandardID(id) |
| 187 |
elif len(id) == 6 and id[4] == '_': |
| 188 |
return SeqResID(id) |
| 189 |
else: |
| 190 |
return DegenerateID(id) |
| 191 |
|
| 192 |
@abstractmethod |
| 193 |
def parse(self, id): |
| 194 |
""" |
| 195 |
Split C{id} into accession number and chain ID. |
| 196 |
|
| 197 |
@param id: PDB identifier |
| 198 |
@type id: str |
| 199 |
@return: (accession, chain) |
| 200 |
@rtype: tuple of str |
| 201 |
|
| 202 |
@raise InvalidEntryIDError: when C{id} is in an inappropriate format |
| 203 |
""" |
| 204 |
pass |
| 205 |
|
| 206 |
def format(self): |
| 207 |
""" |
| 208 |
@return: the identifier in its original format |
| 209 |
@rtype: str |
| 210 |
""" |
| 211 |
return self.entry_id |
| 212 |
|
| 213 |
@property |
| 214 |
def accession(self): |
| 215 |
""" |
| 216 |
Accession number part of the Entry ID |
| 217 |
@rtype: str |
| 218 |
""" |
| 219 |
return self._accession |
| 220 |
|
| 221 |
@property |
| 222 |
def chain(self): |
| 223 |
""" |
| 224 |
Chain ID part of the Entry ID |
| 225 |
@rtype: str |
| 226 |
""" |
| 227 |
return self._chain |
| 228 |
|
| 229 |
@property |
| 230 |
def entry_id(self): |
| 231 |
""" |
| 232 |
Accession number + Chain ID |
| 233 |
@rtype: str |
| 234 |
""" |
| 235 |
return "{0.accession}{0.chain}".format(self) |
| 236 |
|
| 237 |
def __str__(self): |
| 238 |
return self.entry_id |
| 239 |
|
| 240 |
class StandardID(EntryID): |
| 241 |
""" |
| 242 |
Standard PDB ID in the following form: xxxxY, where xxxx is the accession |
| 243 |
number (lower case) and Y is an optional chain identifier. |
| 244 |
""" |
| 245 |
def parse(self, id): |
| 246 |
|
| 247 |
if len(id) not in (4, 5): |
| 248 |
raise InvalidEntryIDError(id) |
| 249 |
|
| 250 |
return (id[:4].lower(), id[4:]) |
| 251 |
|
| 252 |
class DegenerateID(EntryID): |
| 253 |
""" |
| 254 |
Looks like a L{StandardID}, except that the accession number may have |
| 255 |
arbitrary length. |
| 256 |
""" |
| 257 |
def parse(self, id): |
| 258 |
|
| 259 |
if len(id) < 2: |
| 260 |
raise InvalidEntryIDError(id) |
| 261 |
|
| 262 |
return (id[:-1].lower(), id[-1]) |
| 263 |
|
| 264 |
class SeqResID(EntryID): |
| 265 |
""" |
| 266 |
Same as a L{StandardID}, but contains an additional underscore between |
| 267 |
te accession number and the chain identifier. |
| 268 |
""" |
| 269 |
def parse(self, id): |
| 270 |
|
| 271 |
if not (len(id) == 6 and id[4] == '_'): |
| 272 |
raise InvalidEntryIDError(id) |
| 273 |
|
| 274 |
return (id[:4].lower(), id[5:]) |
| 275 |
|
| 276 |
def format(self): |
| 277 |
return "{0.accession}_{0.chain}".format(self) |
| 278 |
|
| 279 |
|
| 280 |
class AbstractStructureParser(object): |
| 281 |
""" |
| 282 |
A base PDB structure format-aware parser. Subclasses must implement the |
| 283 |
internal abstract method C{_parse_header} in order to complete the |
| 284 |
implementation. |
| 285 |
|
| 286 |
@param structure_file: the input PD file to parse |
| 287 |
@type structure_file: str |
| 288 |
@param check_ss: if True, secondary structure errors in the file will cause |
| 289 |
L{SecStructureFormatError} exceptions |
| 290 |
@type check_ss: bool |
| 291 |
@param mapper: residue mapper, used to align ATOM records to SEQRES. |
| 292 |
If None, use the default (L{CombinedResidueMapper}) |
| 293 |
@type mapper: L{AbstractResidueMapper} |
| 294 |
|
| 295 |
@raise IOError: when the input file cannot be found |
| 296 |
""" |
| 297 |
|
| 298 |
__metaclass__ = ABCMeta |
| 299 |
|
| 300 |
@staticmethod |
| 301 |
def create_parser(structure_file, check_ss=False, mapper=None): |
| 302 |
""" |
| 303 |
A StructureParser factory, which instantiates and returns the proper parser |
| 304 |
object based on the contents of the PDB file. |
| 305 |
|
| 306 |
If the file contains a SEQRES section, L{RegularStructureParser} is returned, |
| 307 |
otherwise L{LegacyStructureParser} is instantiated. In the latter case |
| 308 |
LegacyStructureParser will read the sequence data directly from the ATOMs. |
| 309 |
|
| 310 |
@param structure_file: the PDB file to parse |
| 311 |
@type structure_file: str |
| 312 |
@param check_ss: if True, secondary structure errors in the file will cause |
| 313 |
L{SecStructureFormatError} exceptions |
| 314 |
@type check_ss: bool |
| 315 |
@param mapper: residue mapper, used to align ATOM records to SEQRES. |
| 316 |
If None, use the default (L{CombinedResidueMapper}) |
| 317 |
@type mapper: L{AbstractResidueMapper} |
| 318 |
|
| 319 |
@rtype: L{AbstractStructureParser} |
| 320 |
""" |
| 321 |
has_seqres = False |
| 322 |
|
| 323 |
for line in open(structure_file): |
| 324 |
if line.startswith('SEQRES'): |
| 325 |
has_seqres = True |
| 326 |
break |
| 327 |
|
| 328 |
if has_seqres: |
| 329 |
return RegularStructureParser(structure_file, check_ss, mapper) |
| 330 |
else: |
| 331 |
return LegacyStructureParser(structure_file, check_ss, mapper) |
| 332 |
|
| 333 |
def __init__(self, structure_file, check_ss=False, mapper=None): |
| 334 |
|
| 335 |
self._file = None |
| 336 |
self._stream = None |
| 337 |
self._mapper = CombinedResidueMapper() |
| 338 |
self._check_ss = bool(check_ss) |
| 339 |
|
| 340 |
self.filename = structure_file |
| 341 |
if mapper is not None: |
| 342 |
self.mapper = mapper |
| 343 |
|
| 344 |
def __del__(self): |
| 345 |
try: |
| 346 |
self._stream.close() |
| 347 |
except: |
| 348 |
pass |
| 349 |
|
| 350 |
@property |
| 351 |
def mapper(self): |
| 352 |
""" |
| 353 |
Current residue mapping strategy |
| 354 |
@rtype: L{AbstractResidueMapper} |
| 355 |
""" |
| 356 |
return self._mapper |
| 357 |
@mapper.setter |
| 358 |
def mapper(self, value): |
| 359 |
if not isinstance(value, AbstractResidueMapper): |
| 360 |
raise TypeError(value) |
| 361 |
self._mapper = value |
| 362 |
|
| 363 |
@property |
| 364 |
def filename(self): |
| 365 |
""" |
| 366 |
Current input PDB file name |
| 367 |
@rtype: str |
| 368 |
""" |
| 369 |
return self._file |
| 370 |
@filename.setter |
| 371 |
def filename(self, name): |
| 372 |
try: |
| 373 |
stream = open(name) |
| 374 |
except IOError: |
| 375 |
raise IOError('File not found: {0}'.format(name)) |
| 376 |
|
| 377 |
if self._stream: |
| 378 |
try: |
| 379 |
self._stream.close() |
| 380 |
except: |
| 381 |
pass |
| 382 |
self._stream = stream |
| 383 |
self._file = name |
| 384 |
|
| 385 |
def models(self): |
| 386 |
""" |
| 387 |
Find all available model identifiers in the structure. |
| 388 |
|
| 389 |
@return: a list of model IDs |
| 390 |
@rtype: list |
| 391 |
""" |
| 392 |
models = [] |
| 393 |
check = set() |
| 394 |
|
| 395 |
with open(self._file, 'r') as f: |
| 396 |
for line in f: |
| 397 |
if line.startswith('MODEL'): |
| 398 |
model_id = int(line[10:14]) |
| 399 |
if model_id in check: |
| 400 |
raise StructureFormatError('Duplicate model identifier: {0}'.format(model_id)) |
| 401 |
models.append(model_id) |
| 402 |
check.add(model_id) |
| 403 |
|
| 404 |
if len(models) > 0: |
| 405 |
if not(min(check) == 1 and max(check) == len(models)): |
| 406 |
raise StructureFormatError('Non-consecutive model identifier(s) encountered') |
| 407 |
return models |
| 408 |
else: |
| 409 |
return [] |
| 410 |
|
| 411 |
def guess_chain_type(self, residue_labels): |
| 412 |
""" |
| 413 |
Try to guess what is the sequence type of a chunk of PDB |
| 414 |
C{residue_label}s. The list of labels is probed starting from the middle |
| 415 |
first, because PDB chains often contain modified / unknown residues at |
| 416 |
the termini. If none of the probed residues can be used to determine |
| 417 |
chain's type, just give up and return L{SequenceTypes.Unknown}. |
| 418 |
|
| 419 |
@param residue_labels: an iterable of PDB residue labels |
| 420 |
@type residue_labels: iterable |
| 421 |
|
| 422 |
@return: a L{SequenceTypes} enum member |
| 423 |
@rtype: L{csb.core.EnumItem} |
| 424 |
""" |
| 425 |
|
| 426 |
labels = list(residue_labels) |
| 427 |
middle = int(len(labels) / 2) |
| 428 |
|
| 429 |
reordered = labels[middle:] + list(reversed(labels[:middle])) |
| 430 |
|
| 431 |
for label in reordered: |
| 432 |
try: |
| 433 |
type = self.guess_sequence_type(label) |
| 434 |
if type != SequenceTypes.Unknown: |
| 435 |
return type |
| 436 |
|
| 437 |
except UnknownPDBResidueError: |
| 438 |
continue |
| 439 |
|
| 440 |
return SequenceTypes.Unknown |
| 441 |
|
| 442 |
def guess_sequence_type(self, residue_label): |
| 443 |
""" |
| 444 |
Try to guess what is the sequence type of a PDB C{residue_label}. |
| 445 |
|
| 446 |
@param residue_label: a PDB-conforming name of a residue |
| 447 |
@type residue_label: str |
| 448 |
|
| 449 |
@return: a L{SequenceTypes} enum member |
| 450 |
@rtype: L{csb.core.EnumItem} |
| 451 |
|
| 452 |
@raise UnknownPDBResidueError: when there is no such PDB residue name |
| 453 |
in the catalog tables |
| 454 |
""" |
| 455 |
if residue_label in PDB_AMINOACIDS: |
| 456 |
return SequenceTypes.Protein |
| 457 |
elif residue_label in PDB_NUCLEOTIDES: |
| 458 |
return SequenceTypes.NucleicAcid |
| 459 |
else: |
| 460 |
raise UnknownPDBResidueError(residue_label) |
| 461 |
|
| 462 |
def parse_residue(self, residue_label, as_type=None): |
| 463 |
""" |
| 464 |
Try to parse a PDB C{residue_label} and return its closest 'normal' |
| 465 |
string representation. If a sequence type (C{as_type}) is defined, |
| 466 |
guess the alphabet based on that information, otherwise try first to |
| 467 |
parse it as a protein residue. |
| 468 |
|
| 469 |
@param residue_label: a PDB-conforming name of a residue |
| 470 |
@type residue_label: str |
| 471 |
@param as_type: suggest a sequence type (L{SequenceTypes} member) |
| 472 |
@type L{scb.core.EnumItem} |
| 473 |
|
| 474 |
@return: a normalized residue name |
| 475 |
@rtype: str |
| 476 |
|
| 477 |
@raise UnknownPDBResidueError: when there is no such PDB residue name |
| 478 |
in the catalog table(s) |
| 479 |
""" |
| 480 |
if as_type is None: |
| 481 |
as_type = self.guess_sequence_type(residue_label) |
| 482 |
|
| 483 |
try: |
| 484 |
if as_type == SequenceTypes.Protein: |
| 485 |
return PDB_AMINOACIDS[residue_label] |
| 486 |
elif as_type == SequenceTypes.NucleicAcid: |
| 487 |
return PDB_NUCLEOTIDES[residue_label] |
| 488 |
else: |
| 489 |
raise UnknownPDBResidueError(residue_label) |
| 490 |
except KeyError: |
| 491 |
raise UnknownPDBResidueError(residue_label) |
| 492 |
|
| 493 |
def parse_residue_safe(self, residue_label, as_type): |
| 494 |
""" |
| 495 |
Same as C{parse_residue}, but returns UNK/Any instead of raising |
| 496 |
UnknownPDBResidueError. |
| 497 |
|
| 498 |
@param residue_label: a PDB-conforming name of a residue |
| 499 |
@type residue_label: str |
| 500 |
@param as_type: suggest a sequence type (L{SequenceTypes} member) |
| 501 |
@type L{scb.core.EnumItem} |
| 502 |
|
| 503 |
@return: a normalized residue name |
| 504 |
@rtype: str |
| 505 |
""" |
| 506 |
try: |
| 507 |
return self.parse_residue(residue_label, as_type) |
| 508 |
except UnknownPDBResidueError: |
| 509 |
if as_type == SequenceTypes.Protein: |
| 510 |
return repr(SequenceAlphabets.Protein.UNK) |
| 511 |
elif as_type == SequenceTypes.NucleicAcid: |
| 512 |
return repr(SequenceAlphabets.Nucleic.Any) |
| 513 |
else: |
| 514 |
return repr(SequenceAlphabets.Unknown.UNK) |
| 515 |
|
| 516 |
def parse(self, filename=None, model=None): |
| 517 |
|
| 518 |
if filename: |
| 519 |
self.filename = filename |
| 520 |
return self.parse_structure(model) |
| 521 |
|
| 522 |
def parse_structure(self, model=None): |
| 523 |
""" |
| 524 |
Parse and return the L{Structure} with the specified model identifier. |
| 525 |
If no explicit model is specified, parse the first model in the |
| 526 |
structure. |
| 527 |
|
| 528 |
@param model: parse exactly the model with this ID |
| 529 |
@type model: str |
| 530 |
|
| 531 |
@return: object representation of the selected model |
| 532 |
@rtype: L{Structure} |
| 533 |
|
| 534 |
@raise ValueError: When an invalid model ID is specified |
| 535 |
@raise PDBParseError: When the input PDB file suffers from unrecoverable |
| 536 |
corruption. More specialized exceptions will be |
| 537 |
raised depending on the context (see L{PDBParseError}'s |
| 538 |
subclasses). |
| 539 |
""" |
| 540 |
if model is not None: |
| 541 |
model = int(model) |
| 542 |
|
| 543 |
try: |
| 544 |
structure = self._parse_header(model) |
| 545 |
except PDBParseError: |
| 546 |
raise |
| 547 |
except ValueError as ex: |
| 548 |
raise HeaderFormatError("Malformed header: {0}".format(ex)) |
| 549 |
|
| 550 |
self._parse_atoms(structure, model) |
| 551 |
self._parse_ss(structure) |
| 552 |
|
| 553 |
return structure |
| 554 |
|
| 555 |
def parse_models(self, models=()): |
| 556 |
""" |
| 557 |
Parse the specified models in the file and build an L{Ensemble}. |
| 558 |
|
| 559 |
@param models: an iterable object providing model identifiers. |
| 560 |
If not specified, all models will be parsed. |
| 561 |
@type models: tuple |
| 562 |
|
| 563 |
@return: an ensemble with all parsed models |
| 564 |
@rtype: L{Ensemble} |
| 565 |
""" |
| 566 |
|
| 567 |
if not models: |
| 568 |
models = self.models() |
| 569 |
else: |
| 570 |
models = list(map(int, models)) |
| 571 |
|
| 572 |
ensemble = csb.bio.structure.Ensemble() |
| 573 |
|
| 574 |
if len(models) > 0: |
| 575 |
for model_id in models: |
| 576 |
model = self.parse_structure(model_id) |
| 577 |
ensemble.models.append(model) |
| 578 |
else: |
| 579 |
model = self.parse_structure() |
| 580 |
model.model_id = 1 |
| 581 |
ensemble.models.append(model) |
| 582 |
|
| 583 |
return ensemble |
| 584 |
|
| 585 |
def parse_biomolecule(self, number=1, single=False): |
| 586 |
""" |
| 587 |
Parse and return the L{Structure} of the biological unit (quaternary |
| 588 |
structure) as annotated by the REMARK 350 BIOMOLECULE record. |
| 589 |
|
| 590 |
@param number: biomolecule number |
| 591 |
@type number: int |
| 592 |
|
| 593 |
@param single: if True, assign new single-letter chain |
| 594 |
identifiers. If False, assign multi-letter chain identifiers whith a |
| 595 |
number appended to the original identifier, like "A1", "A2", ... |
| 596 |
@type single: bool |
| 597 |
|
| 598 |
@return: structure of biological unit |
| 599 |
@rtype: L{Structure} |
| 600 |
""" |
| 601 |
remarks = self._parse_remarks() |
| 602 |
if 350 not in remarks: |
| 603 |
raise PDBParseError('There is no REMARK 350') |
| 604 |
|
| 605 |
current = 1 |
| 606 |
biomt = {current: {}} |
| 607 |
chains = tuple() |
| 608 |
|
| 609 |
def split(line): |
| 610 |
return [c.strip() for c in line.split(',') if c.strip() != ''] |
| 611 |
|
| 612 |
for line in remarks[350]: |
| 613 |
if line.startswith('BIOMOLECULE:'): |
| 614 |
current = int(line[12:]) |
| 615 |
biomt[current] = {} |
| 616 |
elif line.startswith('APPLY THE FOLLOWING TO CHAINS:'): |
| 617 |
chains = tuple(split(line[30:])) |
| 618 |
elif line.startswith(' AND CHAINS:'): |
| 619 |
chains += tuple(split(line[30:])) |
| 620 |
elif line.startswith(' BIOMT'): |
| 621 |
num = int(line[8:12]) |
| 622 |
vec = line[12:].split() |
| 623 |
vec = list(map(float, vec)) |
| 624 |
biomt[current].setdefault(chains, dict()).setdefault(num, []).extend(vec) |
| 625 |
|
| 626 |
if number not in biomt or len(biomt[number]) == 0: |
| 627 |
raise KeyError('no BIOMOLECULE number {0}'.format(number)) |
| 628 |
|
| 629 |
asu = self.parse_structure() |
| 630 |
structure = csb.bio.structure.Structure('{0}_{1}'.format(asu.accession, number)) |
| 631 |
|
| 632 |
newchainiditer = iter('ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789') |
| 633 |
|
| 634 |
for chains, matrices in biomt[number].items(): |
| 635 |
for num in matrices: |
| 636 |
mat = numpy.array(matrices[num][0:12]).reshape((3,4)) |
| 637 |
R, t = mat[:3,:3], mat[:3,3] |
| 638 |
|
| 639 |
for chain in chains: |
| 640 |
if chain not in asu: |
| 641 |
raise PDBParseError('chain {0} missing'.format(chain)) |
| 642 |
copy = asu[chain].clone() |
| 643 |
copy.transform(R, t) |
| 644 |
if single: |
| 645 |
if len(structure.chains) == 62: |
| 646 |
raise ValueError('too many chains for single=True') |
| 647 |
copy.id = next(newchainiditer) |
| 648 |
else: |
| 649 |
copy.id = '{0}{1}'.format(chain, num) |
| 650 |
structure.chains.append(copy) |
| 651 |
|
| 652 |
return structure |
| 653 |
|
| 654 |
@abstractmethod |
| 655 |
def _parse_header(self, model): |
| 656 |
""" |
| 657 |
An abstract method, which subclasses must use to customize the way the |
| 658 |
PDB header (or the absence of a such) is handled. The implementation |
| 659 |
must rely on reading character data from the current internal |
| 660 |
self._stream and must return a new L{csb.bio.structure.Structure} |
| 661 |
instance with properly initialized header data: accession, model, |
| 662 |
molecule identifiers, chains and residues. This structure object is |
| 663 |
then internally passed to the C{_parse_atoms} hook, responsible for |
| 664 |
attachment of the atoms to the residues in the structure. |
| 665 |
|
| 666 |
@param model: model ID to parse |
| 667 |
@type model: str |
| 668 |
|
| 669 |
@rtype: L{Structure} |
| 670 |
""" |
| 671 |
pass |
| 672 |
|
| 673 |
def _scroll_model(self, model, stream): |
| 674 |
""" |
| 675 |
Scroll the C{stream} to the specified C{model}. |
| 676 |
""" |
| 677 |
|
| 678 |
while True: |
| 679 |
try: |
| 680 |
line = next(stream) |
| 681 |
except StopIteration: |
| 682 |
raise ValueError('No such model {0} in the structure.'.format(model)) |
| 683 |
|
| 684 |
if line.startswith('MODEL'): |
| 685 |
model_id = self._read_model(line) |
| 686 |
if model == model_id: |
| 687 |
return model_id |
| 688 |
|
| 689 |
def _parse_atoms(self, structure, model): |
| 690 |
""" |
| 691 |
Parse the ATOMs from the specified C{model} and attach them to the |
| 692 |
C{structure}. |
| 693 |
|
| 694 |
@param structure: L{Structure} being constructed |
| 695 |
@type structure:L{Structure} |
| 696 |
@param model: model ID to parse |
| 697 |
@type model: str |
| 698 |
""" |
| 699 |
|
| 700 |
structure.model_id = None |
| 701 |
|
| 702 |
chains = set() |
| 703 |
total_chains = len([c for c in structure.items if c.length > 0]) |
| 704 |
|
| 705 |
residues = dict( (chain, []) for chain in structure.chains ) |
| 706 |
seen_residues = dict( (chain, {}) for chain in structure.chains ) |
| 707 |
|
| 708 |
in_ligands = False |
| 709 |
in_atom = False |
| 710 |
read_model = False |
| 711 |
|
| 712 |
self._stream.seek(0) |
| 713 |
while True: |
| 714 |
|
| 715 |
try: |
| 716 |
line = next(self._stream) |
| 717 |
except StopIteration: |
| 718 |
break |
| 719 |
|
| 720 |
if line.startswith('MODEL'): |
| 721 |
if read_model: |
| 722 |
break |
| 723 |
else: |
| 724 |
self._parse_model_line(line, structure, model) |
| 725 |
model = structure.model_id |
| 726 |
read_model = True |
| 727 |
|
| 728 |
elif line.startswith('ATOM') or \ |
| 729 |
(line.startswith('HETATM') and not in_ligands): |
| 730 |
in_atom = True |
| 731 |
|
| 732 |
info = self._parse_atom_line(line, structure) |
| 733 |
chains.add(info.chain) |
| 734 |
|
| 735 |
if info.id not in seen_residues[info.chain]: |
| 736 |
residues[info.chain].append(info) |
| 737 |
seen_residues[info.chain][info.id] = info |
| 738 |
else: |
| 739 |
atom = info.atoms[0] |
| 740 |
seen_residues[info.chain][info.id].atoms.append(atom) |
| 741 |
|
| 742 |
elif in_atom and line.startswith('TER'): |
| 743 |
in_atom = False |
| 744 |
if len(chains) == total_chains: |
| 745 |
in_ligands = True |
| 746 |
|
| 747 |
elif line.startswith('ENDMDL'): |
| 748 |
break |
| 749 |
|
| 750 |
elif line.startswith('END'): |
| 751 |
break |
| 752 |
|
| 753 |
if structure.model_id != model: |
| 754 |
raise ValueError('No such model {0} in the structure.'.format(model)) |
| 755 |
|
| 756 |
self._map_residues(structure, residues) |
| 757 |
|
| 758 |
def _parse_model_line(self, line, structure, model): |
| 759 |
""" |
| 760 |
Handle a new MODEL line. The default implementation will read the model |
| 761 |
ID and attach it to the C{structure}. |
| 762 |
|
| 763 |
@param line: raw string line to parse |
| 764 |
@type line: str |
| 765 |
@param structure: L{Structure} being constructed |
| 766 |
@type structure:L{Structure} |
| 767 |
|
| 768 |
@note: this method may have side effects: scrolls the current stream |
| 769 |
|
| 770 |
@return: read model ID |
| 771 |
@rtype: int |
| 772 |
""" |
| 773 |
if model and model != self._read_model(line): |
| 774 |
self._scroll_model(model, self._stream) |
| 775 |
structure.model_id = model |
| 776 |
else: |
| 777 |
model = self._read_model(line) |
| 778 |
structure.model_id = model |
| 779 |
|
| 780 |
return model |
| 781 |
|
| 782 |
def _parse_atom_line(self, line, structure): |
| 783 |
""" |
| 784 |
Handle a new ATOM or HETATM line. The default implementation will read |
| 785 |
all data fields and create a new L{Atom}. |
| 786 |
|
| 787 |
@param line: raw string line to parse |
| 788 |
@type line: str |
| 789 |
@param structure: L{Structure} being constructed |
| 790 |
@type structure:L{Structure} |
| 791 |
|
| 792 |
@return: newly constructed atom |
| 793 |
@rtype: L{ResidueInfo} |
| 794 |
""" |
| 795 |
|
| 796 |
atom = self._read_atom(line) |
| 797 |
|
| 798 |
rank = self._read_sequence_number(line) |
| 799 |
sequence_number = rank |
| 800 |
insertion_code = self._read_insertion_code(line) |
| 801 |
|
| 802 |
id = str(sequence_number) |
| 803 |
if insertion_code: |
| 804 |
id += insertion_code |
| 805 |
|
| 806 |
chain = self._read_chain_id(line) |
| 807 |
if chain not in structure.chains: |
| 808 |
raise StructureFormatError("Chain {0} is undefined".format(chain)) |
| 809 |
|
| 810 |
type = self._read_residue(line, structure.chains[chain]) |
| 811 |
label = self._read_residue_raw(line) |
| 812 |
|
| 813 |
atom.alternate = self._read_alternate(line) |
| 814 |
atom.occupancy = self._read_occupancy(line) |
| 815 |
atom.bfactor = self._read_bfactor(line) |
| 816 |
atom.charge = self._read_charge(line) |
| 817 |
|
| 818 |
info = ResidueInfo(chain, rank, id, sequence_number, insertion_code, type, label) |
| 819 |
info.atoms = [atom] |
| 820 |
|
| 821 |
return info |
| 822 |
|
| 823 |
def _read_model(self, line): |
| 824 |
""" |
| 825 |
@return: model identifier |
| 826 |
@rtype: int |
| 827 |
""" |
| 828 |
try: |
| 829 |
return int(line[10:14]) |
| 830 |
except ValueError: |
| 831 |
raise StructureFormatError("Invalid model ID: {0}".format(line)) |
| 832 |
|
| 833 |
def _read_atom(self, line): |
| 834 |
""" |
| 835 |
@return: a new atom (serial_number, name, element, vector) |
| 836 |
@rtype: L{Atom} |
| 837 |
""" |
| 838 |
try: |
| 839 |
serial_number = int(line[6:11]) |
| 840 |
name = line[12:16] |
| 841 |
x, y, z = line[30:38], line[38:46], line[46:54] |
| 842 |
vector = numpy.array([float(x), float(y), float(z)]) |
| 843 |
except ValueError as ve: |
| 844 |
raise StructureFormatError("Invalid ATOM line: {0}".format(ve)) |
| 845 |
|
| 846 |
element = self._read_element(line) |
| 847 |
return csb.bio.structure.Atom(serial_number, name, element, vector) |
| 848 |
|
| 849 |
def _read_sequence_number(self, line): |
| 850 |
""" |
| 851 |
@return: PDB sequence number |
| 852 |
@rtype: int |
| 853 |
""" |
| 854 |
try: |
| 855 |
return int(line[22:26]) |
| 856 |
except ValueError: |
| 857 |
raise StructureFormatError("Invalid sequence number") |
| 858 |
|
| 859 |
def _read_insertion_code(self, line): |
| 860 |
""" |
| 861 |
@return: PDB insertion code |
| 862 |
@rtype: str or None |
| 863 |
""" |
| 864 |
code = line[26].strip() |
| 865 |
|
| 866 |
if code: |
| 867 |
return code |
| 868 |
else: |
| 869 |
return None |
| 870 |
|
| 871 |
def _read_chain_id(self, line): |
| 872 |
""" |
| 873 |
@return: chain identifier |
| 874 |
@rtype: str |
| 875 |
""" |
| 876 |
return line[21].strip() |
| 877 |
|
| 878 |
def _read_residue(self, line, chain): |
| 879 |
""" |
| 880 |
@param chain: owning L{Chain} object |
| 881 |
@type chain: L{Chain} |
| 882 |
|
| 883 |
@return: a member of any alphabet (e.g. L{SequenceAlphabets.Protein}) |
| 884 |
@rtype: L{EnumItem} |
| 885 |
""" |
| 886 |
raw = self._read_residue_raw(line) |
| 887 |
residue = self.parse_residue_safe(raw, as_type=chain.type) |
| 888 |
|
| 889 |
try: |
| 890 |
if chain.type == SequenceTypes.NucleicAcid: |
| 891 |
return csb.core.Enum.parsename(SequenceAlphabets.Nucleic, residue) |
| 892 |
else: |
| 893 |
return csb.core.Enum.parsename(SequenceAlphabets.Protein, residue) |
| 894 |
except csb.core.EnumMemberError: |
| 895 |
raise StructureFormatError("{0} is not a valid {1} residue".format(raw, chain.type)) |
| 896 |
|
| 897 |
def _read_residue_raw(self, line): |
| 898 |
""" |
| 899 |
@rtype: str |
| 900 |
""" |
| 901 |
return line[17:20].strip() |
| 902 |
|
| 903 |
def _read_element(self, line): |
| 904 |
""" |
| 905 |
@return: a member of L{ChemElements} |
| 906 |
@rtype: L{EnumItem} or None |
| 907 |
""" |
| 908 |
element = line[76:78].strip() |
| 909 |
if element: |
| 910 |
try: |
| 911 |
element = csb.core.Enum.parsename(ChemElements, element) |
| 912 |
except csb.core.EnumMemberError: |
| 913 |
if element in ('D', 'X'): |
| 914 |
element = ChemElements.x |
| 915 |
else: |
| 916 |
raise StructureFormatError('Unknown chemical element: {0}'.format(element)) |
| 917 |
else: |
| 918 |
element = None |
| 919 |
|
| 920 |
return element |
| 921 |
|
| 922 |
def _read_alternate(self, line): |
| 923 |
""" |
| 924 |
@return: alt identifier |
| 925 |
@rtype: str or None |
| 926 |
""" |
| 927 |
alternate = line[16].strip() |
| 928 |
|
| 929 |
if not alternate: |
| 930 |
return None |
| 931 |
else: |
| 932 |
return alternate |
| 933 |
|
| 934 |
def _read_occupancy(self, line): |
| 935 |
""" |
| 936 |
@return: occupancy |
| 937 |
@rtype: float or None |
| 938 |
""" |
| 939 |
try: |
| 940 |
return float(line[54:60].strip() or 0) |
| 941 |
except ValueError: |
| 942 |
raise StructureFormatError("Malformed occupancy field") |
| 943 |
|
| 944 |
def _read_bfactor(self, line): |
| 945 |
""" |
| 946 |
@return: b-factor |
| 947 |
@rtype: float or None |
| 948 |
""" |
| 949 |
try: |
| 950 |
return float(line[60:66].strip() or 0) |
| 951 |
except ValueError: |
| 952 |
raise StructureFormatError("Malformed bfactor field") |
| 953 |
|
| 954 |
def _read_charge(self, line): |
| 955 |
""" |
| 956 |
@return: charge |
| 957 |
@rtype: int or None |
| 958 |
""" |
| 959 |
charge = line[78:80].strip() |
| 960 |
|
| 961 |
if charge: |
| 962 |
if charge in ('+', '-'): |
| 963 |
charge += '1' |
| 964 |
if charge[-1] in ('+', '-'): |
| 965 |
charge = charge[::-1] |
| 966 |
try: |
| 967 |
return int(charge) |
| 968 |
except ValueError: |
| 969 |
raise StructureFormatError("Malformed charge field") |
| 970 |
else: |
| 971 |
return None |
| 972 |
|
| 973 |
def _map_residues(self, structure, residues): |
| 974 |
""" |
| 975 |
Attach each L{Atom} to its corresponding L{Residue}. |
| 976 |
|
| 977 |
So far we have constructed a sparse (fragmented) chain given the information |
| 978 |
we have read from the ATOM/HETATM records. That includes PDB sequence |
| 979 |
identifiers and insertion codes, which cover only residues with XYZ coordinates |
| 980 |
and often do not correspond to our L{Residue} ranks. |
| 981 |
Our job is to find the right correspondence by matching this unreal sequence |
| 982 |
to what we have got from the SEQRES fields (the complete, gap-free protein |
| 983 |
sequence). Here we delegate this task to the current L{AbstractResidueMapper} |
| 984 |
strategy, which does all the magic. |
| 985 |
|
| 986 |
@param structure: L{Structure} being constructed |
| 987 |
@type structure:L{Structure} |
| 988 |
@param residues: all L{Atom}s which have been constructed so far. |
| 989 |
This must be a map of the form: |
| 990 |
<chainID: [L{ResidueInfo}1, L{ResidueInfo}2...]> |
| 991 |
@type residues: dict of L{ResidueInfo} |
| 992 |
""" |
| 993 |
|
| 994 |
if set(structure.chains) != set(residues.keys()): |
| 995 |
raise PDBParseError("Corrupt PDB file") |
| 996 |
|
| 997 |
for chain in structure.items: |
| 998 |
if chain.length == 0 or len(residues[chain.id]) == 0: |
| 999 |
continue |
| 1000 |
|
| 1001 |
reference = SparseChainSequence.create(chain) |
| 1002 |
sparse = SparseChainSequence( |
| 1003 |
"atoms", "", residues[chain.id], chain.type) |
| 1004 |
|
| 1005 |
aligned = self.mapper.map(sparse, reference) |
| 1006 |
assert aligned.length == chain.length |
| 1007 |
|
| 1008 |
for residue, mapped in zip(chain.residues, aligned.residues): |
| 1009 |
if mapped.type != sparse.alphabet.GAP: |
| 1010 |
residue.id = (mapped.sequence_number, mapped.insertion_code) |
| 1011 |
|
| 1012 |
for atom in mapped.atoms: |
| 1013 |
residue.atoms.append(atom) |
| 1014 |
|
| 1015 |
def _parse_ss(self, structure): |
| 1016 |
""" |
| 1017 |
Parse and attach secondary structure data. |
| 1018 |
|
| 1019 |
@bug: Currently the PDB helix types are ignored. Each HELIX line is treated |
| 1020 |
as a regular SecStructures.Helix. This is due to incompatibility |
| 1021 |
between DSSP and PDB helix types. |
| 1022 |
@todo: Implement a proper workaround for the previous bug (e.g. skip all |
| 1023 |
helices types not included in the DSSP enum) |
| 1024 |
|
| 1025 |
@warning: In this implementation only the start/end positions of the SS |
| 1026 |
elements are parsed. Additional data like H-bonding is ignored. |
| 1027 |
|
| 1028 |
@bug: Currently structure.to_pdb() is not writing any SS data. |
| 1029 |
""" |
| 1030 |
elements = {} |
| 1031 |
self._stream.seek(0) |
| 1032 |
|
| 1033 |
while True: |
| 1034 |
try: |
| 1035 |
line = next(self._stream) |
| 1036 |
except StopIteration: |
| 1037 |
break |
| 1038 |
|
| 1039 |
if line.startswith('HELIX'): |
| 1040 |
|
| 1041 |
chain = structure.chains[line[19].strip()] |
| 1042 |
if chain.id not in elements: |
| 1043 |
elements[chain.id] = [] |
| 1044 |
if chain.id != line[31].strip(): |
| 1045 |
if self._check_ss: |
| 1046 |
raise SecStructureFormatError('Helix {0} spans multiple chains'.format(line[7:10])) |
| 1047 |
else: |
| 1048 |
continue |
| 1049 |
try: |
| 1050 |
startres = chain.find(line[21:25].strip(), line[25].strip()) |
| 1051 |
endres = chain.find(line[33:37].strip(), line[37].strip()) |
| 1052 |
except csb.core.ItemNotFoundError as ex: |
| 1053 |
if self._check_ss: |
| 1054 |
raise SecStructureFormatError( |
| 1055 |
'Helix {0} refers to an undefined residue ID: {1}'.format(line[7:10], str(ex))) |
| 1056 |
else: |
| 1057 |
continue |
| 1058 |
if not startres.rank <= endres.rank: |
| 1059 |
if self._check_ss: |
| 1060 |
raise SecStructureFormatError('Helix {0} is out of range'.format(line[7:10])) |
| 1061 |
else: |
| 1062 |
continue |
| 1063 |
helix = csb.bio.structure.SecondaryStructureElement(startres.rank, endres.rank, SecStructures.Helix) |
| 1064 |
elements[chain.id].append(helix) |
| 1065 |
|
| 1066 |
if line.startswith('SHEET'): |
| 1067 |
|
| 1068 |
chain = structure.chains[line[21].strip()] |
| 1069 |
if chain.id not in elements: |
| 1070 |
elements[chain.id] = [] |
| 1071 |
if chain.id != line[32].strip(): |
| 1072 |
if self._check_ss: |
| 1073 |
raise SecStructureFormatError('Sheet {0} spans multiple chains'.format(line[7:10])) |
| 1074 |
else: |
| 1075 |
continue |
| 1076 |
try: |
| 1077 |
startres = chain.find(line[22:26].strip(), line[26].strip()) |
| 1078 |
endres = chain.find(line[33:37].strip(), line[37].strip()) |
| 1079 |
except csb.core.ItemNotFoundError as ex: |
| 1080 |
if self._check_ss: |
| 1081 |
raise SecStructureFormatError( |
| 1082 |
'Sheet {0} refers to an undefined residue ID: {1}'.format(line[7:10], str(ex))) |
| 1083 |
else: |
| 1084 |
continue |
| 1085 |
if not startres.rank <= endres.rank: |
| 1086 |
if self._check_ss: |
| 1087 |
raise SecStructureFormatError('Sheet {0} is out of range'.format(line[7:10])) |
| 1088 |
else: |
| 1089 |
continue |
| 1090 |
strand = csb.bio.structure.SecondaryStructureElement(startres.rank, endres.rank, SecStructures.Strand) |
| 1091 |
elements[chain.id].append(strand) |
| 1092 |
|
| 1093 |
elif line.startswith('MODEL') or line.startswith('ATOM'): |
| 1094 |
break |
| 1095 |
|
| 1096 |
for chain_id in elements: |
| 1097 |
ss = csb.bio.structure.SecondaryStructure() |
| 1098 |
for e in elements[chain_id]: |
| 1099 |
ss.append(e) |
| 1100 |
structure.chains[chain_id].secondary_structure = ss |
| 1101 |
|
| 1102 |
def _parse_remarks(self): |
| 1103 |
""" |
| 1104 |
Read REMARK lines from PDB file. |
| 1105 |
|
| 1106 |
@return: dictionary with remark numbers as keys, and lists of lines as values. |
| 1107 |
@rtype: dict |
| 1108 |
""" |
| 1109 |
self._stream.seek(0) |
| 1110 |
|
| 1111 |
remarks = {} |
| 1112 |
|
| 1113 |
for line in self._stream: |
| 1114 |
if line.startswith('REMARK'): |
| 1115 |
num = int(line[7:10]) |
| 1116 |
lstring = line[11:] |
| 1117 |
remarks.setdefault(num, []).append(lstring) |
| 1118 |
elif line.startswith('DBREF') or line.startswith('ATOM'): |
| 1119 |
break |
| 1120 |
|
| 1121 |
return remarks |
| 1122 |
|
| 1123 |
|
| 1124 |
class RegularStructureParser(AbstractStructureParser): |
| 1125 |
""" |
| 1126 |
This is the de facto PDB parser, which is designed to read SEQRES and ATOM |
| 1127 |
sections separately, and them map them. Intentionally fails to parse |
| 1128 |
malformed PDB files, e.g. a PDB file without a HEADER section. |
| 1129 |
""" |
| 1130 |
|
| 1131 |
def _parse_header(self, model): |
| 1132 |
""" |
| 1133 |
Parse the HEADER section of a regular PDB file. |
| 1134 |
|
| 1135 |
@return: a L{csb.bio.structure.Structure} instance with properly |
| 1136 |
initialized residues from the SEQRES. |
| 1137 |
@rtype: L{csb.bio.structure.Structure} |
| 1138 |
|
| 1139 |
@raise PDBParseError: if the stream has no HEADER at byte 0 |
| 1140 |
""" |
| 1141 |
|
| 1142 |
self._stream.seek(0) |
| 1143 |
|
| 1144 |
header = next(self._stream) |
| 1145 |
if not header.startswith('HEADER'): |
| 1146 |
raise PDBParseError('Does not look like a regular PDB file.') |
| 1147 |
|
| 1148 |
structure = csb.bio.structure.Structure(header.split()[-1]) |
| 1149 |
|
| 1150 |
while True: |
| 1151 |
|
| 1152 |
try: |
| 1153 |
line = next(self._stream) |
| 1154 |
except StopIteration: |
| 1155 |
break |
| 1156 |
|
| 1157 |
if line.startswith('COMPND'): |
| 1158 |
if line[10:].lstrip().startswith('MOL_ID:'): |
| 1159 |
mol_id = int(line[18:].replace(';', '').strip()) |
| 1160 |
chain_name = '' |
| 1161 |
chains = '' |
| 1162 |
while line.startswith('COMPND'): |
| 1163 |
line = next(self._stream) |
| 1164 |
if line.split()[2].startswith('MOLECULE:'): |
| 1165 |
chain_name += line[20:].strip() |
| 1166 |
while not chain_name.endswith(';'): |
| 1167 |
line = next(self._stream) |
| 1168 |
if not line.startswith('COMPND'): |
| 1169 |
break |
| 1170 |
chain_name += ' ' + line[11:].strip() |
| 1171 |
else: |
| 1172 |
while not line.split()[2].startswith('CHAIN:'): |
| 1173 |
line = next(self._stream) |
| 1174 |
if not line.startswith('COMPND'): |
| 1175 |
raise HeaderFormatError('Missing chain identifier in COMPND section') |
| 1176 |
chains = line[17:].strip() |
| 1177 |
while not chains.endswith(';'): |
| 1178 |
line = next(self._stream) |
| 1179 |
if not line.startswith('COMPND'): |
| 1180 |
break |
| 1181 |
chains += ', ' + line[11:].strip() |
| 1182 |
break |
| 1183 |
|
| 1184 |
chain_ids = chains.replace(';', ' ').replace(',', ' ').split() or [''] # the second part deals with an empty chain id |
| 1185 |
self._add_chains(structure, chain_name, mol_id, *chain_ids) |
| 1186 |
|
| 1187 |
elif line.startswith('REMARK 2 RESOLUTION'): |
| 1188 |
structure.resolution = self._read_resolution(line) |
| 1189 |
|
| 1190 |
elif line.startswith('SEQRES'): |
| 1191 |
chain_id, residues = self._parse_seqres_line(line, structure) |
| 1192 |
chain = structure.chains[chain_id] |
| 1193 |
|
| 1194 |
for residue in residues: |
| 1195 |
chain.residues.append(residue) |
| 1196 |
if chain.residues.last_index != residue.rank: |
| 1197 |
raise HeaderFormatError("Malformed SEQRES") |
| 1198 |
|
| 1199 |
elif line.startswith('MODEL') or line.startswith('ATOM'): |
| 1200 |
break |
| 1201 |
|
| 1202 |
return structure |
| 1203 |
|
| 1204 |
def _add_chains(self, structure, name, mol_id, *chain_ids): |
| 1205 |
|
| 1206 |
name = name.strip().rstrip(";") |
| 1207 |
|
| 1208 |
for chain in chain_ids: |
| 1209 |
new_chain = csb.bio.structure.Chain(chain, type=SequenceTypes.Unknown, |
| 1210 |
name=name, accession=structure.accession) |
| 1211 |
new_chain.molecule_id = mol_id |
| 1212 |
try: |
| 1213 |
structure.chains.append(new_chain) |
| 1214 |
except csb.bio.structure.DuplicateChainIDError: |
| 1215 |
raise HeaderFormatError('Chain {0} is already defined.'.format(new_chain.id)) |
| 1216 |
|
| 1217 |
def _read_resolution(self, line): |
| 1218 |
""" |
| 1219 |
@return: resolution |
| 1220 |
@rtype: float or None |
| 1221 |
""" |
| 1222 |
res = re.search("(\d+(?:\.\d+)?)\s+ANGSTROM", line) |
| 1223 |
|
| 1224 |
if res and res.groups(): |
| 1225 |
return float(res.group(1)) |
| 1226 |
else: |
| 1227 |
return None |
| 1228 |
|
| 1229 |
def _parse_seqres_line(self, line, structure): |
| 1230 |
""" |
| 1231 |
Parse a SEQRES line, build and return newly constructed residues. |
| 1232 |
If the current sequence type of the chain is unknown, try to guess it |
| 1233 |
before parsing the residues. |
| 1234 |
|
| 1235 |
@return: parsed chain_id and L{Residue}s |
| 1236 |
@rtype: 2-tuple: (str, iterable of L{Residue}) |
| 1237 |
""" |
| 1238 |
residues = [] |
| 1239 |
|
| 1240 |
rownum = int(line[7:10]) |
| 1241 |
chain_id = line[11].strip() |
| 1242 |
labels = line[18:].split() |
| 1243 |
|
| 1244 |
if chain_id not in structure.chains: |
| 1245 |
raise HeaderFormatError('Chain {0} is undefined'.format(chain_id)) |
| 1246 |
|
| 1247 |
chain = structure.chains[chain_id] |
| 1248 |
|
| 1249 |
if chain.type == SequenceTypes.Unknown: |
| 1250 |
chain.type = self.guess_chain_type(labels) |
| 1251 |
|
| 1252 |
for rn, label in enumerate(labels): |
| 1253 |
rank = rownum * 13 - (13 - (rn + 1)) |
| 1254 |
rtype = self.parse_residue_safe(label, as_type=chain.type) |
| 1255 |
|
| 1256 |
residue = csb.bio.structure.Residue.create(chain.type, rank=rank, type=rtype) |
| 1257 |
residue.label = label |
| 1258 |
residues.append(residue) |
| 1259 |
|
| 1260 |
return chain_id, residues |
| 1261 |
|
| 1262 |
|
| 1263 |
class PDBHeaderParser(RegularStructureParser): |
| 1264 |
""" |
| 1265 |
Ultra fast PDB HEADER parser. Does not read any structural data. |
| 1266 |
""" |
| 1267 |
|
| 1268 |
def _parse_atoms(self, structure, model): |
| 1269 |
pass |
| 1270 |
|
| 1271 |
def _parse_ss(self, structure): |
| 1272 |
pass |
| 1273 |
|
| 1274 |
def _parse_header(self, model): |
| 1275 |
return super(PDBHeaderParser, self)._parse_header(model) |
| 1276 |
|
| 1277 |
|
| 1278 |
class LegacyStructureParser(AbstractStructureParser): |
| 1279 |
""" |
| 1280 |
This is a customized PDB parser, which is designed to read both sequence and |
| 1281 |
atom data from the ATOM section. This is especially useful when parsing PDB |
| 1282 |
files without a header. |
| 1283 |
""" |
| 1284 |
|
| 1285 |
def _parse_header(self, model): |
| 1286 |
""" |
| 1287 |
Initialize a structure with residues from the ATOMs section. |
| 1288 |
|
| 1289 |
@param model: model identifier (e.g. if multiple models exist) |
| 1290 |
@type model: str |
| 1291 |
|
| 1292 |
@return: a L{csb.bio.structure.Structure} instance with properly |
| 1293 |
initialized residues from ATOMs under the specified C{model}. |
| 1294 |
@rtype: L{csb.bio.structure.Structure} |
| 1295 |
""" |
| 1296 |
|
| 1297 |
self._stream.seek(0) |
| 1298 |
in_atom = False |
| 1299 |
has_atoms = False |
| 1300 |
has_model = False |
| 1301 |
chains = csb.core.OrderedDict() |
| 1302 |
|
| 1303 |
header = next(self._stream) |
| 1304 |
if header.startswith('HEADER'): |
| 1305 |
structure = csb.bio.structure.Structure(header.split()[-1]) |
| 1306 |
else: |
| 1307 |
self._stream.seek(0) |
| 1308 |
structure = csb.bio.structure.Structure('NONE') |
| 1309 |
|
| 1310 |
structure.model_id = None |
| 1311 |
|
| 1312 |
while True: |
| 1313 |
|
| 1314 |
try: |
| 1315 |
line = next(self._stream) |
| 1316 |
except StopIteration: |
| 1317 |
break |
| 1318 |
|
| 1319 |
if line.startswith('MODEL'): |
| 1320 |
if has_model: |
| 1321 |
break |
| 1322 |
else: |
| 1323 |
self._parse_model_line(line, structure, model) |
| 1324 |
model = structure.model_id |
| 1325 |
has_model = True |
| 1326 |
|
| 1327 |
elif line.startswith('ATOM') \ |
| 1328 |
or (in_atom and line.startswith('HETATM')): |
| 1329 |
in_atom = True |
| 1330 |
has_atoms = True |
| 1331 |
|
| 1332 |
seq_number = self._read_sequence_number(line) |
| 1333 |
ins_code = self._read_insertion_code(line) |
| 1334 |
residue_id = (seq_number, ins_code) |
| 1335 |
label = self._read_residue_raw(line) |
| 1336 |
chain_id = self._read_chain_id(line) |
| 1337 |
|
| 1338 |
if chain_id not in chains: |
| 1339 |
chains[chain_id] = csb.core.OrderedDict() |
| 1340 |
self._add_chain(structure, chain_id) |
| 1341 |
|
| 1342 |
if residue_id not in chains[chain_id]: |
| 1343 |
chains[chain_id][residue_id] = label |
| 1344 |
chain = structure.chains[chain_id] |
| 1345 |
if chain.type == SequenceTypes.Unknown: |
| 1346 |
self._fix_chain(chain, label) |
| 1347 |
|
| 1348 |
elif in_atom and line.startswith('TER'): |
| 1349 |
in_atom = False |
| 1350 |
elif line.startswith('ENDMDL'): |
| 1351 |
break |
| 1352 |
elif line.startswith('END'): |
| 1353 |
break |
| 1354 |
|
| 1355 |
if not has_atoms: |
| 1356 |
raise HeaderFormatError("Can't parse legacy structure: no ATOMs found") |
| 1357 |
|
| 1358 |
for chain in structure.items: |
| 1359 |
self._build_chain(chain, chains[chain.id]) |
| 1360 |
|
| 1361 |
return structure |
| 1362 |
|
| 1363 |
def _add_chain(self, structure, chain_id): |
| 1364 |
|
| 1365 |
new_chain = csb.bio.structure.Chain(chain_id, |
| 1366 |
type=SequenceTypes.Unknown, |
| 1367 |
accession=structure.accession) |
| 1368 |
new_chain.molecule_id = '1' |
| 1369 |
structure.chains.append(new_chain) |
| 1370 |
|
| 1371 |
def _build_chain(self, chain, residues): |
| 1372 |
|
| 1373 |
for residue_id, label in residues.items(): |
| 1374 |
rank = (chain.residues.last_index or 0) + 1 |
| 1375 |
|
| 1376 |
rname = self.parse_residue_safe(label, as_type=chain.type) |
| 1377 |
residue = csb.bio.structure.Residue.create(chain.type, rank=rank, type=rname) |
| 1378 |
residue.label = label |
| 1379 |
residue.id = residue_id |
| 1380 |
chain.residues.append(residue) |
| 1381 |
|
| 1382 |
def _fix_chain(self, chain, probe): |
| 1383 |
|
| 1384 |
try: |
| 1385 |
chain.type = self.guess_sequence_type(probe) |
| 1386 |
except UnknownPDBResidueError: |
| 1387 |
pass |
| 1388 |
|
| 1389 |
def _map_residues(self, structure, residues): |
| 1390 |
|
| 1391 |
for chain in structure.items: |
| 1392 |
for residue_info in residues[chain.id]: |
| 1393 |
try: |
| 1394 |
residue = chain.find(residue_info.sequence_number, residue_info.insertion_code) |
| 1395 |
for atom in residue_info.atoms: |
| 1396 |
residue.atoms.append(atom) |
| 1397 |
|
| 1398 |
except csb.bio.structure.EntityNotFoundError: |
| 1399 |
pass |
| 1400 |
|
| 1401 |
|
| 1402 |
StructureParser = AbstractStructureParser.create_parser |
| 1403 |
""" |
| 1404 |
Alias for L{AbstractStructureParser.create_parser}. |
| 1405 |
""" |
| 1406 |
|
| 1407 |
|
| 1408 |
class ResidueInfo(object): |
| 1409 |
""" |
| 1410 |
High-performance struct, which functions as a container for unmapped |
| 1411 |
L{Atom}s. |
| 1412 |
|
| 1413 |
@note: This object must implement the L{csb.bio.sequence.ResidueInfo} |
| 1414 |
interface. This is not enforced through inheritance solely |
| 1415 |
to save some CPU (by exposing public fields and no properties). |
| 1416 |
However, on an abstract level this object is_a ResidueInfo |
| 1417 |
and is used to build L{AbstractSequence}s. |
| 1418 |
""" |
| 1419 |
__slots__ = ['chain', 'rank', 'id' , 'sequence_number', 'insertion_code', 'type', 'label', 'atoms'] |
| 1420 |
|
| 1421 |
def __init__(self, chain, rank, id, seq_number, ins_code, type, label): |
| 1422 |
|
| 1423 |
self.chain = chain |
| 1424 |
self.rank = rank |
| 1425 |
self.id = id |
| 1426 |
self.sequence_number = seq_number |
| 1427 |
self.insertion_code = ins_code |
| 1428 |
self.type = type |
| 1429 |
self.label = label |
| 1430 |
self.atoms = [] |
| 1431 |
|
| 1432 |
@property |
| 1433 |
def is_modified(self): |
| 1434 |
|
| 1435 |
if self.type.enum is SequenceAlphabets.Nucleic: |
| 1436 |
return self.label != str(self.type) |
| 1437 |
else: |
| 1438 |
return self.label != repr(self.type) |
| 1439 |
|
| 1440 |
class SparseChainSequence(csb.bio.sequence.ChainSequence): |
| 1441 |
""" |
| 1442 |
Sequence view for reference (SEQRES) or sparse (ATOM) PDB chains. |
| 1443 |
The residue instances passed to the constructor must be |
| 1444 |
L{csb.bio.structure.Residue} or L{csb.bio.io.wwpdb.ResidueInfo} objects. |
| 1445 |
|
| 1446 |
See L{csb.bio.sequence.AbstractSequence} for details. |
| 1447 |
""" |
| 1448 |
|
| 1449 |
def _add(self, residue): |
| 1450 |
|
| 1451 |
if not isinstance(residue, (csb.bio.structure.Residue, ResidueInfo)): |
| 1452 |
raise TypeError(residue) |
| 1453 |
else: |
| 1454 |
self._residues.append(residue) |
| 1455 |
|
| 1456 |
def _get(self, rank): |
| 1457 |
return self._residues[rank - 1] |
| 1458 |
|
| 1459 |
@staticmethod |
| 1460 |
def create(chain): |
| 1461 |
""" |
| 1462 |
Create a new L{SparseChainSequence} from existing L{Chain}. |
| 1463 |
|
| 1464 |
@type chain: L{csb.bio.structure.Chain} |
| 1465 |
@rtype: L{SparseChainSequence} |
| 1466 |
""" |
| 1467 |
return SparseChainSequence( |
| 1468 |
chain.entry_id, chain.header, chain.residues, chain.type) |
| 1469 |
|
| 1470 |
|
| 1471 |
class AbstractResidueMapper(object): |
| 1472 |
""" |
| 1473 |
Defines the base interface of all residue mappers, used to align PDB ATOM |
| 1474 |
records to the real (SEQRES) sequence of a chain. |
| 1475 |
""" |
| 1476 |
|
| 1477 |
__metaclass__ = ABCMeta |
| 1478 |
|
| 1479 |
@abstractmethod |
| 1480 |
def map(self, sparse, reference): |
| 1481 |
""" |
| 1482 |
Map L{sparse}'s residues to L{reference}. Return all C{sparse} residues, |
| 1483 |
aligned over C{reference}, with artificial gap residues inserted at |
| 1484 |
relevant positions. The resulting sequence of sparse residues will |
| 1485 |
always have the same length as the C{reference} sequence. |
| 1486 |
|
| 1487 |
@note: C{sparse}'s ranks won't be touched because the C{rank} property |
| 1488 |
of the underlying L{ResidueInfo} implementation is not necessarily r/w. |
| 1489 |
|
| 1490 |
@param sparse: sparse sequence (e.g. derived from ATOMS records) |
| 1491 |
@type sparse: L{SparseChainSequence} |
| 1492 |
@param reference: reference, complete sequence |
| 1493 |
(e.g. derived from SEQRES records) |
| 1494 |
@type reference: L{SparseChainSequence} |
| 1495 |
|
| 1496 |
@return: all C{sparse} residues, optimally aligned over C{reference} |
| 1497 |
(with gaps) |
| 1498 |
@rtype: L{SparseChainSequence} |
| 1499 |
|
| 1500 |
@raise ResidueMappingError: if the specified sequences are not alignable |
| 1501 |
""" |
| 1502 |
pass |
| 1503 |
|
| 1504 |
def create_gap(self, alphabet=SequenceAlphabets.Protein): |
| 1505 |
""" |
| 1506 |
Create and return a new gap residue. |
| 1507 |
|
| 1508 |
@param alphabet: sequence alphabet; a member of L{SequenceAlphabets} |
| 1509 |
which has GAP item |
| 1510 |
@type alphabet: L{enum} |
| 1511 |
|
| 1512 |
@rtype: L{ResidueInfo} |
| 1513 |
""" |
| 1514 |
return ResidueInfo(None, -1, None, None, None, alphabet.GAP, "-") |
| 1515 |
|
| 1516 |
def _build(self, sparse, aligned): |
| 1517 |
|
| 1518 |
return SparseChainSequence( |
| 1519 |
sparse.id, sparse.header, aligned, sparse.type) |
| 1520 |
|
| 1521 |
class FastResidueMapper(AbstractResidueMapper): |
| 1522 |
""" |
| 1523 |
RegExp-based residue mapper. Fails on heavily malformed input (i.e. it cannot |
| 1524 |
insert gaps in the C{reference}), but it is very fast (linear) and memory |
| 1525 |
efficient. |
| 1526 |
""" |
| 1527 |
|
| 1528 |
MAX_FRAGMENTS = 20 |
| 1529 |
|
| 1530 |
MIN_UNICODE_CHAR = 300 |
| 1531 |
FORBIDDEN_CHARS = set('^.*?()-') |
| 1532 |
|
| 1533 |
CODEC = "utf-8" |
| 1534 |
DELIMITER = ").*?(".encode(CODEC).decode(CODEC) |
| 1535 |
PATTERN = "^.*?({0}).*?$".encode(CODEC).decode(CODEC) |
| 1536 |
|
| 1537 |
def __init__(self): |
| 1538 |
self._charcode = FastResidueMapper.MIN_UNICODE_CHAR |
| 1539 |
self._cache = {} |
| 1540 |
|
| 1541 |
def map(self, sparse, reference): |
| 1542 |
|
| 1543 |
aligned = [] |
| 1544 |
mapping = {} |
| 1545 |
|
| 1546 |
residues = list(sparse.residues) |
| 1547 |
|
| 1548 |
pattern = self._build_pattern(residues) |
| 1549 |
seqres = self._encode_sequence(reference) |
| 1550 |
|
| 1551 |
matches = re.match(pattern, seqres) |
| 1552 |
|
| 1553 |
if matches: |
| 1554 |
unmapped_item = -1 |
| 1555 |
|
| 1556 |
for fn, fragment in enumerate(matches.groups(), start=1): |
| 1557 |
assert fragment != '' |
| 1558 |
|
| 1559 |
for offset in range(1, len(fragment) + 1): |
| 1560 |
unmapped_item += 1 |
| 1561 |
rank = matches.start(fn) + offset |
| 1562 |
|
| 1563 |
mapped_residue = residues[unmapped_item] |
| 1564 |
real_residue = reference.residues[rank] |
| 1565 |
|
| 1566 |
assert real_residue.type == mapped_residue.type |
| 1567 |
mapping[real_residue] = mapped_residue |
| 1568 |
|
| 1569 |
else: |
| 1570 |
raise ResidueMappingError("Can't map ATOM records") |
| 1571 |
|
| 1572 |
for rank, residue in enumerate(reference.residues, start=1): |
| 1573 |
if residue in mapping: |
| 1574 |
aligned.append(mapping[residue]) |
| 1575 |
else: |
| 1576 |
aligned.append(self.create_gap(sparse.alphabet)) |
| 1577 |
|
| 1578 |
assert len(aligned) == reference.length |
| 1579 |
return self._build(sparse, aligned) |
| 1580 |
|
| 1581 |
def _build_pattern(self, residues): |
| 1582 |
""" |
| 1583 |
Build and return a sparse regular rexpression for C{residues}. |
| 1584 |
""" |
| 1585 |
|
| 1586 |
fragments = [] |
| 1587 |
|
| 1588 |
for rn, r in enumerate(residues): |
| 1589 |
res_name = self._encode(r) |
| 1590 |
|
| 1591 |
if rn == 0: |
| 1592 |
# First residue, start a new fragment: |
| 1593 |
fragments.append([res_name]) |
| 1594 |
elif r.insertion_code: # and not residues[rn - 1].insertion_code: |
| 1595 |
# If residue i has an insertion code, initiate a new fragment: |
| 1596 |
fragments.append([res_name]) |
| 1597 |
elif r.sequence_number - residues[rn - 1].sequence_number in (0, 1, -1): |
| 1598 |
# If the seq numbers of residues [i-1, i] are consecutive, extend the last fragment: |
| 1599 |
fragments[-1].append(res_name) |
| 1600 |
else: |
| 1601 |
# They are not consecutive, so we better start a new fragment: |
| 1602 |
fragments.append([res_name]) |
| 1603 |
|
| 1604 |
for i, frag in enumerate(fragments): |
| 1605 |
fragments[i] = ''.join(frag) |
| 1606 |
if len(fragments) > FastResidueMapper.MAX_FRAGMENTS: |
| 1607 |
# Wow, that's a lot of fragments. Better use a different mapper |
| 1608 |
raise ResidueMappingError("Can't map chain with large number of fragments") |
| 1609 |
|
| 1610 |
blocks = FastResidueMapper.DELIMITER.join(fragments) |
| 1611 |
pattern = FastResidueMapper.PATTERN.format(blocks) |
| 1612 |
|
| 1613 |
return pattern |
| 1614 |
|
| 1615 |
def _encode(self, r): |
| 1616 |
""" |
| 1617 |
Return a unique single-letter representation of C{r.type}. |
| 1618 |
""" |
| 1619 |
if not r.is_modified: |
| 1620 |
return str(r.type) |
| 1621 |
else: |
| 1622 |
return self._register_label(r.label) |
| 1623 |
|
| 1624 |
def _encode_sequence(self, s): |
| 1625 |
return ''.join(map(self._encode, s.residues)) |
| 1626 |
|
| 1627 |
def _register_label(self, label): |
| 1628 |
""" |
| 1629 |
Assign a new unicode character to C{label} and cache it. |
| 1630 |
|
| 1631 |
@return: cached single-letter representation of label. |
| 1632 |
@rtype: unicode char |
| 1633 |
""" |
| 1634 |
|
| 1635 |
if label not in self._cache: |
| 1636 |
if set(label).intersection(FastResidueMapper.FORBIDDEN_CHARS): |
| 1637 |
raise ResidueMappingError("Invalid residue label") |
| 1638 |
|
| 1639 |
self._charcode += 1 |
| 1640 |
code = self._charcode |
| 1641 |
self._cache[label] = csb.io.unichr(code) |
| 1642 |
|
| 1643 |
return self._cache[label] |
| 1644 |
|
| 1645 |
|
| 1646 |
class RobustResidueMapper(AbstractResidueMapper): |
| 1647 |
""" |
| 1648 |
Exhaustive residue mapper, which uses Needleman-Wunsch global alignment. |
| 1649 |
Much slower (quadratic), but fail-proof even with incompatible sequences |
| 1650 |
(can insert gaps in both the C{sparse} and the C{reference} sequence). |
| 1651 |
|
| 1652 |
@param match: score for a match |
| 1653 |
@type match: float |
| 1654 |
@param mismatch: score for a mismatch (by default mismatches are heavily |
| 1655 |
penalized, while gaps are allowed) |
| 1656 |
@type mismatch: float |
| 1657 |
@param gap: gap penalty |
| 1658 |
@type gap: float |
| 1659 |
""" |
| 1660 |
|
| 1661 |
class GlobalAligner(alignment.GlobalAlignmentAlgorithm): |
| 1662 |
|
| 1663 |
def _sequence(self, s): |
| 1664 |
return [r.label for r in s.residues] |
| 1665 |
|
| 1666 |
|
| 1667 |
def __init__(self, match=1, mismatch=-10, gap=0): |
| 1668 |
|
| 1669 |
scoring = alignment.IdentityMatrix(match=match, mismatch=mismatch) |
| 1670 |
aligner = RobustResidueMapper.GlobalAligner(scoring=scoring, gap=gap) |
| 1671 |
|
| 1672 |
self._aligner = aligner |
| 1673 |
|
| 1674 |
def map(self, sparse, reference): |
| 1675 |
|
| 1676 |
aligned = [] |
| 1677 |
ali = self._aligner.align(sparse, reference) |
| 1678 |
|
| 1679 |
if ali.is_empty: |
| 1680 |
raise ResidueMappingError("Global alignment failed") |
| 1681 |
|
| 1682 |
for mapped, residue in zip(ali.query, ali.subject): |
| 1683 |
|
| 1684 |
if residue.type == reference.alphabet.GAP: |
| 1685 |
continue |
| 1686 |
elif mapped.type == sparse.alphabet.GAP: |
| 1687 |
aligned.append(self.create_gap(sparse.alphabet)) |
| 1688 |
else: |
| 1689 |
aligned.append(mapped) |
| 1690 |
|
| 1691 |
return self._build(sparse, aligned) |
| 1692 |
|
| 1693 |
class CombinedResidueMapper(AbstractResidueMapper): |
| 1694 |
""" |
| 1695 |
The best of both worlds: attempts to map the residues using |
| 1696 |
L{FastResidueMapper}, but upon failure secures success by switching to |
| 1697 |
L{RobustResidueMapper}. |
| 1698 |
""" |
| 1699 |
FAST = FastResidueMapper() |
| 1700 |
ROBUST = RobustResidueMapper() |
| 1701 |
|
| 1702 |
def map(self, sparse, reference): |
| 1703 |
|
| 1704 |
try: |
| 1705 |
return CombinedResidueMapper.FAST.map(sparse, reference) |
| 1706 |
except ResidueMappingError: |
| 1707 |
return CombinedResidueMapper.ROBUST.map(sparse, reference) |
| 1708 |
|
| 1709 |
|
| 1710 |
class FileBuilder(object): |
| 1711 |
""" |
| 1712 |
Base abstract files for all structure file formatters. |
| 1713 |
Defines a common step-wise interface according to the Builder pattern. |
| 1714 |
|
| 1715 |
@param output: output stream (this is where the product is constructed) |
| 1716 |
@type output: stream |
| 1717 |
""" |
| 1718 |
|
| 1719 |
__metaclass__ = ABCMeta |
| 1720 |
|
| 1721 |
def __init__(self, output): |
| 1722 |
|
| 1723 |
if not hasattr(output, 'write'): |
| 1724 |
raise TypeError(output) |
| 1725 |
|
| 1726 |
def isnull(this, that, null=None): |
| 1727 |
if this is null: |
| 1728 |
return that |
| 1729 |
else: |
| 1730 |
return this |
| 1731 |
|
| 1732 |
self._out = output |
| 1733 |
self._isnull = isnull |
| 1734 |
|
| 1735 |
@property |
| 1736 |
def output(self): |
| 1737 |
""" |
| 1738 |
Destination stream |
| 1739 |
@rtype: stream |
| 1740 |
""" |
| 1741 |
return self._out |
| 1742 |
|
| 1743 |
@property |
| 1744 |
def isnull(self): |
| 1745 |
""" |
| 1746 |
ISNULL(X, Y) function |
| 1747 |
@rtype: callable |
| 1748 |
""" |
| 1749 |
return self._isnull |
| 1750 |
|
| 1751 |
def write(self, text): |
| 1752 |
""" |
| 1753 |
Write a chunk of text |
| 1754 |
""" |
| 1755 |
self._out.write(text) |
| 1756 |
|
| 1757 |
def writeline(self, text): |
| 1758 |
""" |
| 1759 |
Write a chunk of text and append a new line terminator |
| 1760 |
""" |
| 1761 |
self._out.write(text) |
| 1762 |
self._out.write('\n') |
| 1763 |
|
| 1764 |
@abstractmethod |
| 1765 |
def add_header(self, master_structure): |
| 1766 |
pass |
| 1767 |
|
| 1768 |
@abstractmethod |
| 1769 |
def add_structure(self, structure): |
| 1770 |
pass |
| 1771 |
|
| 1772 |
def finalize(self): |
| 1773 |
pass |
| 1774 |
|
| 1775 |
class PDBFileBuilder(FileBuilder): |
| 1776 |
""" |
| 1777 |
PDB file format builder. |
| 1778 |
""" |
| 1779 |
|
| 1780 |
def writeline(self, text): |
| 1781 |
self.write('{0:80}\n'.format(text)) |
| 1782 |
|
| 1783 |
def add_header(self, master): |
| 1784 |
""" |
| 1785 |
Write the HEADER of the file using C{master} |
| 1786 |
|
| 1787 |
@type master: L{Structure} |
| 1788 |
""" |
| 1789 |
|
| 1790 |
isnull = self.isnull |
| 1791 |
|
| 1792 |
header = 'HEADER {0:40}{1:%d-%b-%y} {2:4}' |
| 1793 |
self.writeline(header.format('.', datetime.datetime.now(), master.accession.upper())) |
| 1794 |
|
| 1795 |
molecules = { } |
| 1796 |
|
| 1797 |
for chain_id in master.chains: |
| 1798 |
chain = master.chains[chain_id] |
| 1799 |
if chain.molecule_id not in molecules: |
| 1800 |
molecules[chain.molecule_id] = [ ] |
| 1801 |
molecules[chain.molecule_id].append(chain_id) |
| 1802 |
|
| 1803 |
k = 0 |
| 1804 |
for mol_id in sorted(molecules): |
| 1805 |
|
| 1806 |
chains = molecules[mol_id] |
| 1807 |
first_chain = master.chains[ chains[0] ] |
| 1808 |
|
| 1809 |
self.writeline('COMPND {0:3} MOL_ID: {1};'.format(k + 1, isnull(mol_id, '0'))) |
| 1810 |
self.writeline('COMPND {0:3} MOLECULE: {1};'.format(k + 2, isnull(first_chain.name, ''))) |
| 1811 |
self.writeline('COMPND {0:3} CHAIN: {1};'.format(k + 3, ', '.join(chains))) |
| 1812 |
k += 3 |
| 1813 |
|
| 1814 |
for chain_id in master.chains: |
| 1815 |
|
| 1816 |
chain = master.chains[chain_id] |
| 1817 |
res = [ r.label for r in chain.residues ] |
| 1818 |
|
| 1819 |
rn = 0 |
| 1820 |
for j in range(0, chain.length, 13): |
| 1821 |
rn += 1 |
| 1822 |
residues = [ '{0:>3}'.format(r) for r in res[j : j + 13] ] |
| 1823 |
self.writeline('SEQRES {0:>3} {1} {2:>4} {3}'.format( |
| 1824 |
rn, chain.id, chain.length, ' '.join(residues) )) |
| 1825 |
|
| 1826 |
def add_structure(self, structure): |
| 1827 |
""" |
| 1828 |
Append a new model to the file |
| 1829 |
|
| 1830 |
@type structure: L{Structure} |
| 1831 |
""" |
| 1832 |
|
| 1833 |
isnull = self.isnull |
| 1834 |
|
| 1835 |
for chain_id in structure.chains: |
| 1836 |
|
| 1837 |
chain = structure.chains[chain_id] |
| 1838 |
for residue in chain.residues: |
| 1839 |
|
| 1840 |
atoms = [ ] |
| 1841 |
for an in residue.atoms: |
| 1842 |
atom = residue.atoms[an] |
| 1843 |
if isinstance(atom, csb.bio.structure.DisorderedAtom): |
| 1844 |
for dis_atom in atom: atoms.append(dis_atom) |
| 1845 |
else: |
| 1846 |
atoms.append(atom) |
| 1847 |
atoms.sort() |
| 1848 |
|
| 1849 |
for atom in atoms: |
| 1850 |
|
| 1851 |
alt = atom.alternate |
| 1852 |
if alt is True: |
| 1853 |
alt = 'A' |
| 1854 |
elif alt is False: |
| 1855 |
alt = ' ' |
| 1856 |
|
| 1857 |
if atom.element: |
| 1858 |
element = repr(atom.element) |
| 1859 |
else: |
| 1860 |
element = ' ' |
| 1861 |
self.writeline('ATOM {0:>5} {1:>4}{2}{3:>3} {4}{5:>4}{6} {7:>8.3f}{8:>8.3f}{9:>8.3f}{10:>6.2f}{11:>6.2f}{12:>12}{13:2}'.format( |
| 1862 |
atom.serial_number, atom._full_name, isnull(alt, ' '), |
| 1863 |
residue.label, chain.id, |
| 1864 |
isnull(residue.sequence_number, residue.rank), isnull(residue.insertion_code, ' '), |
| 1865 |
atom.vector[0], atom.vector[1], atom.vector[2], isnull(atom.occupancy, 0.0), isnull(atom.bfactor, 0.0), |
| 1866 |
element, isnull(atom.charge, ' ') )) |
| 1867 |
|
| 1868 |
self.writeline('TER') |
| 1869 |
|
| 1870 |
def finalize(self): |
| 1871 |
""" |
| 1872 |
Add the END marker |
| 1873 |
""" |
| 1874 |
self.writeline('END') |
| 1875 |
self._out.flush() |
| 1876 |
|
| 1877 |
class PDBEnsembleFileBuilder(PDBFileBuilder): |
| 1878 |
""" |
| 1879 |
Supports serialization of NMR ensembles. |
| 1880 |
|
| 1881 |
Functions as a simple decorator, which wraps C{add_structure} with |
| 1882 |
MODEL/ENDMDL records. |
| 1883 |
""" |
| 1884 |
|
| 1885 |
def add_structure(self, structure): |
| 1886 |
|
| 1887 |
model_id = self.isnull(structure.model_id, 1) |
| 1888 |
|
| 1889 |
self.writeline('MODEL {0:>4}'.format(model_id)) |
| 1890 |
super(PDBEnsembleFileBuilder, self).add_structure(structure) |
| 1891 |
self.writeline('ENDMDL') |
| 1892 |
|
| 1893 |
|
| 1894 |
class StructureProvider(object): |
| 1895 |
""" |
| 1896 |
Base class for all PDB data source providers. |
| 1897 |
|
| 1898 |
Concrete classes need to implement the C{find} method, which abstracts the |
| 1899 |
retrieval of a PDB structure file by a structure identifier. This is a hook |
| 1900 |
method called internally by C{get}, but subclasses can safely override both |
| 1901 |
C{find} and {get} to in order to achieve completely custom behavior. |
| 1902 |
""" |
| 1903 |
|
| 1904 |
__metaclass__ = ABCMeta |
| 1905 |
|
| 1906 |
def __getitem__(self, id): |
| 1907 |
return self.get(id) |
| 1908 |
|
| 1909 |
@abstractmethod |
| 1910 |
def find(self, id): |
| 1911 |
""" |
| 1912 |
Attempt to discover a PDB file, given a specific PDB C{id}. |
| 1913 |
|
| 1914 |
@param id: structure identifier (e.g. 1x80) |
| 1915 |
@type id: str |
| 1916 |
@return: path and file name on success, None otherwise |
| 1917 |
@rtype: str or None |
| 1918 |
""" |
| 1919 |
pass |
| 1920 |
|
| 1921 |
def get(self, id, model=None): |
| 1922 |
""" |
| 1923 |
Discover, parse and return the PDB structure, corresponding to the |
| 1924 |
specified C{id}. |
| 1925 |
|
| 1926 |
@param id: structure identifier (e.g. 1x80) |
| 1927 |
@type id: str |
| 1928 |
@param model: optional model identifier |
| 1929 |
@type model: str |
| 1930 |
|
| 1931 |
@rtype: L{csb.bio.Structure} |
| 1932 |
|
| 1933 |
@raise StructureNotFoundError: when C{id} could not be found |
| 1934 |
""" |
| 1935 |
pdb = self.find(id) |
| 1936 |
|
| 1937 |
if pdb is None: |
| 1938 |
raise StructureNotFoundError(id) |
| 1939 |
else: |
| 1940 |
return StructureParser(pdb).parse_structure(model=model) |
| 1941 |
|
| 1942 |
class FileSystemStructureProvider(StructureProvider): |
| 1943 |
""" |
| 1944 |
Simple file system based PDB data source. Scans a list of local directories |
| 1945 |
using pre-defined file name templates. |
| 1946 |
|
| 1947 |
@param paths: a list of paths |
| 1948 |
@type paths: iterable or str |
| 1949 |
""" |
| 1950 |
|
| 1951 |
def __init__(self, paths=None): |
| 1952 |
|
| 1953 |
self._templates = ['pdb{id}.ent', 'pdb{id}.pdb', '{id}.pdb', '{id}.ent'] |
| 1954 |
self._paths = csb.core.OrderedDict() |
| 1955 |
|
| 1956 |
if paths is not None: |
| 1957 |
if isinstance(paths, csb.core.string): |
| 1958 |
paths = [paths] |
| 1959 |
|
| 1960 |
for path in paths: |
| 1961 |
self.add(path) |
| 1962 |
|
| 1963 |
@property |
| 1964 |
def paths(self): |
| 1965 |
""" |
| 1966 |
Current search paths |
| 1967 |
@rtype: tuple |
| 1968 |
""" |
| 1969 |
return tuple(self._paths) |
| 1970 |
|
| 1971 |
@property |
| 1972 |
def templates(self): |
| 1973 |
""" |
| 1974 |
Current file name match templates |
| 1975 |
@rtype: tuple |
| 1976 |
""" |
| 1977 |
return tuple(self._templates) |
| 1978 |
|
| 1979 |
def add(self, path): |
| 1980 |
""" |
| 1981 |
Register a new local C{path}. |
| 1982 |
|
| 1983 |
@param path: directory name |
| 1984 |
@type path: str |
| 1985 |
|
| 1986 |
@raise IOError: if C{path} is not a valid directory |
| 1987 |
""" |
| 1988 |
if os.path.isdir(path): |
| 1989 |
self._paths[path] = path |
| 1990 |
else: |
| 1991 |
raise IOError(path) |
| 1992 |
|
| 1993 |
def add_template(self, template): |
| 1994 |
""" |
| 1995 |
Register a custom file name name C{template}. The template must contain |
| 1996 |
an E{lb}idE{rb} macro, e.g. pdbE{lb}idE{rb}.ent |
| 1997 |
|
| 1998 |
@param template: pattern |
| 1999 |
@type template: str |
| 2000 |
""" |
| 2001 |
|
| 2002 |
if '{id}' not in template: |
| 2003 |
raise ValueError('Template does not contain an "{id}" macro') |
| 2004 |
|
| 2005 |
if template not in self._templates: |
| 2006 |
self._templates.append(template) |
| 2007 |
|
| 2008 |
def remove(self, path): |
| 2009 |
""" |
| 2010 |
Unregister an existing local C{path}. |
| 2011 |
|
| 2012 |
@param path: directory name |
| 2013 |
@type path: str |
| 2014 |
|
| 2015 |
@raise ValueError: if C{path} had not been registered |
| 2016 |
""" |
| 2017 |
if path not in self._paths: |
| 2018 |
raise ValueError('path not found: {0}'.format(path)) |
| 2019 |
|
| 2020 |
del self._paths[path] |
| 2021 |
|
| 2022 |
def find(self, id): |
| 2023 |
|
| 2024 |
for path in self._paths: |
| 2025 |
for token in self.templates: |
| 2026 |
fn = os.path.join(path, token.format(id=id)) |
| 2027 |
if os.path.exists(fn): |
| 2028 |
return fn |
| 2029 |
|
| 2030 |
return None |
| 2031 |
|
| 2032 |
class RemoteStructureProvider(StructureProvider): |
| 2033 |
""" |
| 2034 |
Retrieves PDB structures from a specified remote URL. |
| 2035 |
The URL requested from remote server takes the form: <prefix>/<ID><suffix> |
| 2036 |
|
| 2037 |
@param prefix: URL prefix, including protocol |
| 2038 |
@type prefix: str |
| 2039 |
@param suffix: optional URL suffix (.ent by default) |
| 2040 |
@type suffix: str |
| 2041 |
""" |
| 2042 |
|
| 2043 |
def __init__(self, prefix='https://files.rcsb.org/download/', suffix='.pdb'): |
| 2044 |
|
| 2045 |
self._prefix = None |
| 2046 |
self._suffix = None |
| 2047 |
|
| 2048 |
self.prefix = prefix |
| 2049 |
self.suffix = suffix |
| 2050 |
|
| 2051 |
@property |
| 2052 |
def prefix(self): |
| 2053 |
""" |
| 2054 |
Current URL prefix |
| 2055 |
@rtype: str |
| 2056 |
""" |
| 2057 |
return self._prefix |
| 2058 |
@prefix.setter |
| 2059 |
def prefix(self, value): |
| 2060 |
self._prefix = value |
| 2061 |
|
| 2062 |
@property |
| 2063 |
def suffix(self): |
| 2064 |
""" |
| 2065 |
Current URL suffix |
| 2066 |
@rtype: str |
| 2067 |
""" |
| 2068 |
return self._suffix |
| 2069 |
@suffix.setter |
| 2070 |
def suffix(self, value): |
| 2071 |
self._suffix = value |
| 2072 |
|
| 2073 |
def _find(self, id): |
| 2074 |
|
| 2075 |
try: |
| 2076 |
return csb.io.urllib.urlopen(self.prefix + id + self.suffix) |
| 2077 |
except Exception as e: |
| 2078 |
raise StructureNotFoundError(id + ": " + str(e)) |
| 2079 |
|
| 2080 |
def find(self, id): |
| 2081 |
|
| 2082 |
stream = self._find(id) |
| 2083 |
|
| 2084 |
try: |
| 2085 |
tmp = csb.io.TempFile(dispose=False) |
| 2086 |
tmp.write(stream.read().decode('utf-8')) |
| 2087 |
tmp.flush() |
| 2088 |
return tmp.name |
| 2089 |
|
| 2090 |
except StructureNotFoundError: |
| 2091 |
return None |
| 2092 |
finally: |
| 2093 |
stream.close() |
| 2094 |
|
| 2095 |
def get(self, id, model=None): |
| 2096 |
|
| 2097 |
stream = self._find(id) |
| 2098 |
|
| 2099 |
try: |
| 2100 |
with csb.io.TempFile() as tmp: |
| 2101 |
tmp.write(stream.read().decode('utf-8')) |
| 2102 |
tmp.flush() |
| 2103 |
return StructureParser(tmp.name).parse_structure(model=model) |
| 2104 |
finally: |
| 2105 |
stream.close() |
| 2106 |
|
| 2107 |
class CustomStructureProvider(StructureProvider): |
| 2108 |
""" |
| 2109 |
A custom PDB data source. Functions as a user-defined map of structure |
| 2110 |
identifiers and their corresponding local file names. |
| 2111 |
|
| 2112 |
@param files: initialization dictionary of id:file pairs |
| 2113 |
@type files: dict-like |
| 2114 |
""" |
| 2115 |
|
| 2116 |
def __init__(self, files={}): |
| 2117 |
|
| 2118 |
self._files = {} |
| 2119 |
for id in files: |
| 2120 |
self.add(id, files[id]) |
| 2121 |
|
| 2122 |
@property |
| 2123 |
def paths(self): |
| 2124 |
""" |
| 2125 |
List of currently registered file names |
| 2126 |
@rtype: tuple |
| 2127 |
""" |
| 2128 |
return tuple(self._files.values()) |
| 2129 |
|
| 2130 |
@property |
| 2131 |
def identifiers(self): |
| 2132 |
""" |
| 2133 |
List of currently registered structure identifiers |
| 2134 |
@rtype: tuple |
| 2135 |
""" |
| 2136 |
return tuple(self._files) |
| 2137 |
|
| 2138 |
def add(self, id, path): |
| 2139 |
""" |
| 2140 |
Register a new local C{id}:C{path} pair. |
| 2141 |
|
| 2142 |
@param id: structure identifier |
| 2143 |
@type id: str |
| 2144 |
@param path: path and file name |
| 2145 |
@type path: str |
| 2146 |
|
| 2147 |
@raise IOError: if C{path} is not a valid file name |
| 2148 |
""" |
| 2149 |
if os.path.isfile(path): |
| 2150 |
self._files[id] = path |
| 2151 |
else: |
| 2152 |
raise IOError(path) |
| 2153 |
|
| 2154 |
def remove(self, id): |
| 2155 |
""" |
| 2156 |
Unregister an existing structure C{id}. |
| 2157 |
|
| 2158 |
@param id: structure identifier |
| 2159 |
@type id: str |
| 2160 |
|
| 2161 |
@raise ValueError: if C{id} had not been registered |
| 2162 |
""" |
| 2163 |
if id not in self._files: |
| 2164 |
raise ValueError(id) |
| 2165 |
else: |
| 2166 |
del self._files[id] |
| 2167 |
|
| 2168 |
def find(self, id): |
| 2169 |
|
| 2170 |
if id in self._files: |
| 2171 |
return self._files[id] |
| 2172 |
else: |
| 2173 |
return None |
| 2174 |
|
| 2175 |
def get(accession, model=None, prefix='https://files.rcsb.org/download/'): |
| 2176 |
""" |
| 2177 |
Download and parse a PDB entry. |
| 2178 |
|
| 2179 |
@param accession: accession number of the entry |
| 2180 |
@type accession: str |
| 2181 |
@param model: model identifier |
| 2182 |
@type model: str |
| 2183 |
@param prefix: download URL prefix |
| 2184 |
@type prefix: str |
| 2185 |
|
| 2186 |
@return: object representation of the selected model |
| 2187 |
@rtype: L{Structure} |
| 2188 |
""" |
| 2189 |
return RemoteStructureProvider(prefix).get(accession, model=model) |
| 2190 |
|
| 2191 |
def find(id, paths): |
| 2192 |
""" |
| 2193 |
Try to discover a PDB file for PDB C{id} in C{paths}. |
| 2194 |
|
| 2195 |
@param id: PDB ID of the entry |
| 2196 |
@type id: str |
| 2197 |
@param paths: a list of directories to scan |
| 2198 |
@type paths: list of str |
| 2199 |
|
| 2200 |
@return: path and file name on success, None otherwise |
| 2201 |
@rtype: str |
| 2202 |
""" |
| 2203 |
return FileSystemStructureProvider(paths).find(id) |
| 2204 |
|
| 2205 |
|
| 2206 |
class AsyncParseResult(object): |
| 2207 |
|
| 2208 |
def __init__(self, result, exception): |
| 2209 |
|
| 2210 |
self.result = result |
| 2211 |
self.exception = exception |
| 2212 |
|
| 2213 |
def __repr__(self): |
| 2214 |
return '<AsyncParseResult: result={0.result}, error={0.exception.__class__.__name__}>'.format(self) |
| 2215 |
|
| 2216 |
|
| 2217 |
def _parse_async(parser, file, model): |
| 2218 |
p = parser(file) |
| 2219 |
return p.parse_structure(model) |
| 2220 |
|
| 2221 |
|
| 2222 |
class AsyncStructureParser(object): |
| 2223 |
""" |
| 2224 |
Wraps StructureParser in an asynchronous call. Since a new process is |
| 2225 |
started by Python internally (as opposed to only starting a new thread), |
| 2226 |
this makes the parser slower, but provides a way to set a parse timeout |
| 2227 |
limit. |
| 2228 |
|
| 2229 |
If initialized with more than one worker, supports parallel parsing |
| 2230 |
through the C{self.parse_async} method. |
| 2231 |
|
| 2232 |
@param workers: number of worker threads (1 by default) |
| 2233 |
@type workers: int |
| 2234 |
""" |
| 2235 |
|
| 2236 |
def __init__(self, workers=1): |
| 2237 |
|
| 2238 |
self._pool = None |
| 2239 |
self._workers = 1 |
| 2240 |
|
| 2241 |
if int(workers) > 0: |
| 2242 |
self._workers = int(workers) |
| 2243 |
else: |
| 2244 |
raise ValueError(workers) |
| 2245 |
|
| 2246 |
self._recycle() |
| 2247 |
|
| 2248 |
def _recycle(self): |
| 2249 |
|
| 2250 |
if self._pool: |
| 2251 |
self._pool.terminate() |
| 2252 |
|
| 2253 |
self._pool = multiprocessing.Pool(processes=self._workers) |
| 2254 |
|
| 2255 |
def parse_structure(self, structure_file, timeout, model=None, |
| 2256 |
parser=RegularStructureParser): |
| 2257 |
""" |
| 2258 |
Call StructureParser.parse_structure() in a separate process and return |
| 2259 |
the output. Raise TimeoutError if the parser does not respond within |
| 2260 |
C{timeout} seconds. |
| 2261 |
|
| 2262 |
@param structure_file: structure file to parse |
| 2263 |
@type structure_file: str |
| 2264 |
@param timeout: raise multiprocessing.TimeoutError if C{timeout} seconds |
| 2265 |
elapse before the parser completes its job |
| 2266 |
@type timeout: int |
| 2267 |
@param parser: any implementing L{AbstractStructureParser} class |
| 2268 |
@type parser: type |
| 2269 |
|
| 2270 |
@return: parsed structure |
| 2271 |
@rtype: L{csb.structure.Structure} |
| 2272 |
""" |
| 2273 |
|
| 2274 |
r = self.parse_async([structure_file], timeout, model, parser) |
| 2275 |
if len(r) > 0: |
| 2276 |
if r[0].exception is not None: |
| 2277 |
raise r[0].exception |
| 2278 |
else: |
| 2279 |
return r[0].result |
| 2280 |
return None |
| 2281 |
|
| 2282 |
def parse_async(self, structure_files, timeout, model=None, |
| 2283 |
parser=RegularStructureParser): |
| 2284 |
""" |
| 2285 |
Call C{self.parse_structure} for a list of structure files |
| 2286 |
simultaneously. The actual degree of parallelism will depend on the |
| 2287 |
number of workers specified while constructing the parser object. |
| 2288 |
|
| 2289 |
@note: Don't be tempted to pass a large list of structures to this |
| 2290 |
method. Every time a C{TimeoutError} is encountered, the |
| 2291 |
corresponding worker process in the pool will hang until the |
| 2292 |
process terminates on its own. During that time, this worker is |
| 2293 |
unusable. If a sufficiently high number of timeouts occur, the |
| 2294 |
whole pool of workers will be unsable. At the end of the method |
| 2295 |
however a pool cleanup is performed and any unusable workers |
| 2296 |
are 'reactivated'. However, that only happens at B{the end} of |
| 2297 |
C{parse_async}. |
| 2298 |
|
| 2299 |
@param structure_files: a list of structure files |
| 2300 |
@type structure_files: tuple of str |
| 2301 |
@param timeout: raise multiprocessing.TimeoutError if C{timeout} seconds |
| 2302 |
elapse before the parser completes its job |
| 2303 |
@type timeout: int |
| 2304 |
@param parser: any implementing L{AbstractStructureParser} class |
| 2305 |
@type parser: type |
| 2306 |
|
| 2307 |
@return: a list of L{AsyncParseResult} objects |
| 2308 |
@rtype: list |
| 2309 |
""" |
| 2310 |
|
| 2311 |
pool = self._pool |
| 2312 |
workers = [] |
| 2313 |
results = [] |
| 2314 |
|
| 2315 |
for file in list(structure_files): |
| 2316 |
result = pool.apply_async(_parse_async, [parser, file, model]) |
| 2317 |
workers.append(result) |
| 2318 |
|
| 2319 |
hanging = False |
| 2320 |
for w in workers: |
| 2321 |
result = AsyncParseResult(None, None) |
| 2322 |
try: |
| 2323 |
result.result = w.get(timeout=timeout) |
| 2324 |
except KeyboardInterrupt as ki: |
| 2325 |
pool.terminate() |
| 2326 |
raise ki |
| 2327 |
except Exception as ex: |
| 2328 |
result.exception = ex |
| 2329 |
if isinstance(ex, multiprocessing.TimeoutError): |
| 2330 |
hanging = True |
| 2331 |
results.append(result) |
| 2332 |
|
| 2333 |
if hanging: |
| 2334 |
self._recycle() |
| 2335 |
|
| 2336 |
return results |