def resegment_reads(fast5_path, params, speedy=False, overwrite=False):
"""Re-segment and create anchor alignment from previously base-called fast5 file
:param fast5_path: path to fast5 file
:param params: event detection parameters
:param speedy: boolean option for speedyStatSplit or minknow
:param overwrite: overwrite a previous event re-segmented event table
:param name: name of key where events table will be placed (Analyses/'name'/Events)
:return True when completed
"""
assert os.path.isfile(fast5_path), "File does not exist: {}".format(fast5_path)
name = "ReSegmentBasecall_00{}"
# create Fast5 object
f5fh = Fast5(fast5_path, read='r+')
# gather previous event detection
old_event_table = f5fh.get_basecall_data()
# assert check_event_table_time(old_event_table), "Old event is not consistent"
read_id = bytes.decode(f5fh.raw_attributes['read_id'])
sampling_freq = f5fh.sample_rate
start_time = f5fh.raw_attributes['start_time']
# pick event detection algorithm
signal = f5fh.get_read(raw=True, scale=True)
if speedy:
event_table = create_speedy_event_table(signal, sampling_freq, start_time, **params)
params = merge_dicts([params, {"event_detection": "speedy_stat_split"}])
else:
event_table = create_minknow_event_table(signal, sampling_freq, start_time, **params)
params = merge_dicts([params, {"event_detection": "minknow_event_detect"}])
keys = ["nanotensor version", "time_stamp"]
values = ["0.2.0", TimeStamp().posix_date()]
attributes = merge_dicts([params, dict(zip(keys, values)), f5fh.raw_attributes])
if f5fh.is_read_rna():
old_event_table = index_to_time(old_event_table, sampling_freq=sampling_freq, start_time=start_time)
# set event table
new_event_table = create_anchor_kmers(new_events=event_table, old_events=old_event_table)
f5fh.set_new_event_table(name, new_event_table, attributes, overwrite=overwrite)
# gather new sequence
sequence = sequence_from_events(new_event_table)
if f5fh.is_read_rna():
sequence = ReverseComplement().reverse(sequence)
sequence = sequence.replace("T", "U")
quality_scores = '!'*len(sequence)
fastq = create_fastq_line(read_id+" :", sequence, quality_scores)
# set fastq
f5fh.set_fastq(name, fastq)
return f5fh
def resegment_reads(fast5_path,
params=None,
speedy=False,
overwrite=True,
analysis_path="ReSegmentBasecall_000"):
"""Re-segment and create anchor alignment from previously base-called fast5 file
:param fast5_path: path to fast5 file
:param params: event detection parameters
:param speedy: boolean option for speedyStatSplit or minknow
:param overwrite: overwrite a previous event re-segmented event table
:param analysis_path: name of key where events table will be placed (Analyses/'name'/Events)
:return True when completed
"""
assert os.path.isfile(fast5_path), "File does not exist: {}".format(
fast5_path)
# create Fast5 object and sanity check
f5fh = Fast5(fast5_path, read='r+')
if not f5fh.has_basecall_data():
f5fh.close()
return None
# gather previous event detection
old_event_table = f5fh.get_basecall_data()
read_id = bytes.decode(f5fh.raw_attributes['read_id'])
sampling_freq = f5fh.sample_rate
start_time = f5fh.raw_attributes['start_time']
# get params
if params is None:
params = get_default_event_detection_params(
EVENT_DETECT_SPEEDY if speedy else EVENT_DETECT_MINKNOW)
# pick event detection algorithm
signal = f5fh.get_read(raw=True, scale=True)
if speedy:
event_table = create_speedy_event_table(signal, sampling_freq,
start_time, **params)
params = merge_dicts(
[params, {
"event_detection": "speedy_stat_split"
}])
else:
event_table = create_minknow_event_table(signal, sampling_freq,
start_time, **params)
params = merge_dicts(
[params, {
"event_detection": "minknow_event_detect"
}])
# metadata
keys = ["nanotensor version", "time_stamp"]
values = ["0.2.0", TimeStamp().posix_date()]
attributes = merge_dicts(
[params, dict(zip(keys, values)), f5fh.raw_attributes])
# do resegmentation
if f5fh.is_read_rna():
old_event_table = index_to_time(old_event_table,
sampling_freq=sampling_freq,
start_time=start_time)
new_event_table = create_anchor_kmers(new_events=event_table,
old_events=old_event_table)
# get destination in fast5
#todo find latest location? ie: save_event_table_and_fastq(..)
destination = f5fh._join_path(f5fh.__base_analysis__, analysis_path)
f5fh.set_event_table(destination,
new_event_table,
attributes,
overwrite=overwrite)
# gather new sequence
sequence = sequence_from_events(new_event_table)
if f5fh.is_read_rna():
sequence = ReverseComplement().reverse(sequence)
sequence = sequence.replace("T", "U")
quality_scores = '!' * len(sequence)
fastq = create_fastq_line(read_id + " :", sequence, quality_scores)
# set fastq
f5fh.set_fastq(destination, fastq, overwrite=overwrite)
return f5fh