def taskmain(predictions_path, prefix_len=0, suffix_len=0,
count=1, random_only=False, n_processes=5, out=None):
start_time = time.time()
predictions = prediction_tools.prepare_model_file(predictions_path)
decoy_path = (predictions_path if out is None else out).rsplit("scored", 1)[0] + "decoy.db"
metadata = predictions.metadata
metadata = copy.deepcopy(metadata)
metadata["tag"] = (metadata["tag"] if metadata["tag"] not in ["", None] else "") + "decoy"
metadata["decoy_ratio"] = count
decoy_logger.info("Saving metadata")
metadata_store = sqlitedict.SqliteDict(decoy_path, tablename="metadata", flag='n')
metadata_store.update(metadata)
metadata_store.commit()
enzyme = predictions.metadata.get("enzyme")
starts = sum([enz["cleavage_start"] for enz in enzyme], [])
ends = sum([enz["cleavage_end"] for enz in enzyme], [])
decoy_logger.info("Building random sequence builder")
decoy_logger.info("Parameters: {}".format(dict(prefix_len=prefix_len,
suffix_len=suffix_len, count=count,
random_only=random_only)))
modification_table = modification.ModificationTable.bootstrap()
builder = RandomGlycopeptideBuilder(
ppm_error=predictions.metadata.get("ms1_ppm_tolerance", 10e-6),
constant_modifications=map(modification_table.__getitem__,
predictions.metadata.get("constant_modifications", [])),
variable_modifications=map(modification_table.__getitem__,
predictions.metadata.get("variable_modifciations", [])),
glycans=glycans_from_predictions(predictions),
cleavage_start=starts,
cleavage_end=ends)
task_fn = functools.partial(build_shuffle_sequences, prefix_len=prefix_len,
suffix_len=suffix_len, count=count,
builder=builder, random_only=random_only)
decoy_logger.info("Beginning generation")
decoy_sequence_store = sqlitedict.SqliteDict(decoy_path,
tablename="theoretical_search_space", flag='w')
total_missing = 0
preds_missed = []
pool = None
if n_processes > 1:
pool = multiprocessing.Pool(n_processes)
dispatcher = functools.partial(pool.imap_unordered, chunksize=1000)
decoy_logger.info("Building decoys concurrently on %d processes", n_processes)
else:
decoy_logger.info("Building decoys sequentially")
dispatcher = itertools.imap
i = 0
pred_iter = 0
for decoy_tuple in (
itertools.chain.from_iterable(
dispatcher(task_fn, predictions.iterrows()))):
decoys, missing = decoy_tuple
for decoy in decoys:
decoy_sequence_store[i] = decoy
i += 1
if i % 10000 == 0:
decoy_sequence_store.commit()
total_missing += missing
if missing > 0:
decoy_logger.info("%d missing", missing)
preds_missed.append((predictions.ix[pred_iter].Glycopeptide_identifier, missing))
pred_iter += 1
if pool is not None:
pool.close()
pool.join()
decoy_sequence_store.commit()
decoy_sequence_store.close()
metadata_store["total_missing_decoys"] = total_missing
metadata_store["predictions_missing_decoys"] = preds_missed
metadata_store.commit()
metadata_store.close()
decoy_logger.info("Done. %d Sequences missing. %f seconds elapsed",
total_missing, time.time() - start_time)
return decoy_path
from glycresoft_ms2_classification import prediction_tools
from glycresoft_ms2_classification.ms import default_loader
predictions = prediction_tools.prepare_model_file("./ResultOf20131219_006_isos.scored_fdr.json").query("MS2_Score >= 0.5")
observed = default_loader("./20131219_006.mzML.20150125-173417matched.pkl").index_by_scan_ids()
spec = observed[predictions.iloc[2].scan_id_range[0]]
drop_mass = predictions.iloc[2].glycanMass
