def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if argv is None:
argv = sys.argv
# setup command line parser
parser = U.OptionParser(version="%prog version: $Id$",
usage=usage,
description=globals()["__doc__"])
group = U.OptionGroup(parser, "whitelist-specific options")
group.add_option("--plot-prefix",
dest="plot_prefix", type="string",
help=("Prefix for plots to visualise the automated "
"detection of the number of 'true' cell barcodes"))
group.add_option("--subset-reads",
dest="subset_reads", type="int",
help=("Use the first N reads to automatically identify "
"the true cell barcodes. If N is greater than the "
"number of reads, all reads will be used. "
"Default is 100,000,000"))
group.add_option("--error-correct-threshold",
dest="error_correct_threshold",
type="int",
help=("Hamming distance for correction of barcodes to "
"whitelist barcodes. This value will also be used "
"for error detection above the knee if required "
"(--ed-above-threshold)"))
group.add_option("--method",
dest="method",
choices=["reads", "umis"],
help=("Use reads or unique umi counts per cell"))
group.add_option("--knee-method",
dest="knee_method",
choices=["distance", "density"],
help=("Use distance or density methods for detection of knee"))
group.add_option("--expect-cells",
dest="expect_cells",
type="int",
help=("Prior expectation on the upper limit on the "
"number of cells sequenced"))
group.add_option("--allow-threshold-error",
dest="allow_threshold_error", action="store_true",
help=("Don't select a threshold. Will still "
"output the plots if requested (--plot-prefix)"))
group.add_option("--set-cell-number",
dest="cell_number",
type="int",
help=("Specify the number of cell barcodes to accept"))
parser.add_option("--ed-above-threshold",
dest="ed_above_threshold", type="choice",
choices=["discard", "correct"],
help=("Detect CBs above the threshold which may be "
"sequence errors from another CB and either "
"'discard' or 'correct'. Default=discard"))
parser.add_option_group(group)
parser.set_defaults(method="reads",
knee_method="distance",
extract_method="string",
whitelist_tsv=None,
blacklist_tsv=None,
error_correct_threshold=1,
pattern=None,
pattern2=None,
read2_in=None,
plot_prefix=None,
subset_reads=100000000,
expect_cells=False,
allow_threshold_error=False,
cell_number=False,
ed_above_threshold=None,
ignore_suffix=False)
# add common options (-h/--help, ...) and parse command line
(options, args) = U.Start(parser, argv=argv,
add_extract_options=True,
add_group_dedup_options=False,
add_umi_grouping_options=False,
add_sam_options=False)
if options.filtered_out and not options.extract_method == "regex":
U.error("Reads will not be filtered unless extract method is"
"set to regex (--extract-method=regex)")
if options.expect_cells:
if options.knee_method == "distance":
U.error("Cannot use --expect-cells with 'distance' knee "
"method. Switch to --knee-method=density if you want to "
"provide an expectation for the number of "
"cells. Alternatively, if you know the number of cell "
"barcodes, use --cell-number")
if options.cell_number:
U.error("Cannot supply both --expect-cells and "
"--cell-number options")
extract_cell, extract_umi = U.validateExtractOptions(options)
if not extract_cell:
if options.extract_method == "string":
U.error("barcode pattern(s) do not include any cell bases "
"(marked with 'Cs') %s, %s" % (
options.pattern, options.pattern2))
elif options.extract_method == "regex":
U.error("barcode regex(es) do not include any cell groups "
"(starting with 'cell_') %s, %s" (
options.pattern, options.pattern2))
read1s = umi_methods.fastqIterate(options.stdin)
# set up read extractor
ReadExtractor = extract_methods.ExtractFilterAndUpdate(
method=options.extract_method,
pattern=options.pattern,
pattern2=options.pattern2,
prime3=options.prime3,
extract_cell=extract_cell)
cell_barcode_counts = collections.Counter()
n_reads = 0
n_cell_barcodes = 0
# if using the umis method, need to keep a set of umis observed
if options.method == "umis":
cell_barcode_umis = collections.defaultdict(set)
# variables for progress monitor
displayMax = 100000
U.info("Starting barcode extraction")
if options.filtered_out:
filtered_out = U.openFile(options.filtered_out, "w")
if not options.read2_in:
for read1 in read1s:
# Update display in every 100kth iteration
if n_reads % displayMax == 0:
U.info("Parsed {} reads".format(n_reads))
n_reads += 1
barcode_values = ReadExtractor.getBarcodes(read1)
if barcode_values is None:
if options.filtered_out:
filtered_out.write(str(read1) + "\n")
continue
else:
cell, umi, _, _, _, _, _ = barcode_values
if options.method == "umis":
cell_barcode_umis[cell].add(umi)
else:
cell_barcode_counts[cell] += 1
n_cell_barcodes += 1
if options.subset_reads:
if n_cell_barcodes > options.subset_reads:
break
else:
if options.filtered_out2:
filtered_out2 = U.openFile(options.filtered_out2, "w")
read2s = umi_methods.fastqIterate(U.openFile(options.read2_in))
for read1, read2 in izip(read1s, read2s):
# Update display in every 100kth iteration
if n_reads % displayMax == 0:
U.info("Parsed {} reads".format(n_reads))
n_reads += 1
barcode_values = ReadExtractor.getBarcodes(read1, read2)
if barcode_values is None:
if options.filtered_out:
filtered_out.write(str(read1) + "\n")
if options.filtered_out2:
filtered_out2.write(str(read2) + "\n")
continue
else:
cell, umi, _, _, _, _, _ = barcode_values
if options.method == "umis":
cell_barcode_umis[cell].add(umi)
else:
cell_barcode_counts[cell] += 1
n_cell_barcodes += 1
if options.subset_reads:
if n_reads > options.subset_reads:
break
U.info("Starting - whitelist determination")
if options.method == "umis":
for cell in cell_barcode_umis:
cell_barcode_counts[cell] = len(cell_barcode_umis[cell])
if options.cell_number and options.cell_number > len(cell_barcode_counts):
raise ValueError(
"--set-cell-barcode option specifies more cell barcodes than the "
"number of observed cell barcodes. This may be because "
"--subset-reads was set to a value too low to capture reads from "
"all cells. %s cell barcodes observed from %s parsed reads. "
"Expected>= %s cell barcodes" % (
len(cell_barcode_counts),
options.subset_reads,
options.cell_number))
cell_whitelist, true_to_false_map = whitelist_methods.getCellWhitelist(
cell_barcode_counts,
options.knee_method,
options.expect_cells,
options.cell_number,
options.error_correct_threshold,
options.plot_prefix)
if cell_whitelist:
U.info("Top %s cell barcodes passed the selected threshold" %
len(cell_whitelist))
if options.ed_above_threshold:
cell_whitelist, true_to_false_map = whitelist_methods.errorDetectAboveThreshold(
cell_barcode_counts,
cell_whitelist,
true_to_false_map,
errors=options.error_correct_threshold,
resolution_method=options.ed_above_threshold)
if cell_whitelist:
U.info("Writing out whitelist")
total_correct_barcodes = 0
total_corrected_barcodes = 0
for barcode in sorted(list(cell_whitelist)):
total_correct_barcodes += cell_barcode_counts[barcode]
if true_to_false_map:
corrected_barcodes = ",".join(
sorted(true_to_false_map[barcode]))
correct_barcode_counts = [cell_barcode_counts[x] for x in
sorted(true_to_false_map[barcode])]
total_corrected_barcodes += sum(correct_barcode_counts)
corrected_barcode_counts = ",".join(
map(str, correct_barcode_counts))
else:
corrected_barcodes, corrected_barcode_counts = "", ""
options.stdout.write("%s\t%s\t%s\t%s\n" % (
barcode, corrected_barcodes, cell_barcode_counts[barcode],
corrected_barcode_counts))
else:
msg = ("No local minima was accepted. Recommend checking the plot "
"output and counts per local minima (requires `--plot-prefix`"
"option) and then re-running with manually selected threshold "
"(`--set-cell-number` option)")
if options.allow_threshold_error:
U.info(msg)
else:
U.error(msg)
U.info("Parsed %i reads" % n_reads)
U.info("%i reads matched the barcode pattern" % n_cell_barcodes)
U.info("Found %i unique cell barcodes" % len(cell_barcode_counts))
if cell_whitelist:
U.info("Found %i total reads matching the selected cell barcodes" %
total_correct_barcodes)
U.info("Found %i total reads which can be error corrected to the "
"selected cell barcodes" % total_corrected_barcodes)
if options.filtered_out:
filtered_out.close()
if options.filtered_out2:
filtered_out2.close()
U.Stop()
def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if argv is None:
argv = sys.argv
# setup command line parser
parser = U.OptionParser(version="%prog version: $Id$",
usage=usage,
description=globals()["__doc__"])
group = U.OptionGroup(parser, "extract-specific options")
# (Experimental option) Retain the UMI in the sequence read"
group.add_option("--retain-umi", dest="retain_umi", action="store_true",
help=optparse.SUPPRESS_HELP)
group.add_option("--read2-out", dest="read2_out", type="string",
help="file to output processed paired read to")
group.add_option("--read2-stdout", dest="read2_stdout",
action="store_true",
help="Paired reads, send read2 to stdout, discarding read1")
group.add_option("--quality-filter-threshold",
dest="quality_filter_threshold", type="int",
help=("Remove reads where any UMI base quality score "
"falls below this threshold"))
group.add_option("--quality-filter-mask",
dest="quality_filter_mask", type="int",
help=("If a UMI base has a quality below this threshold, "
"replace the base with 'N'"))
group.add_option("--quality-encoding",
dest="quality_encoding", type="choice",
choices=["phred33", "phred64", "solexa"],
help=("Quality score encoding. Choose from 'phred33'"
"[33-77] 'phred64' [64-106] or 'solexa' [59-106]"))
group.add_option("--filter-cell-barcode",
dest="filter_cell_barcode",
action="store_true",
help=optparse.SUPPRESS_HELP)
group.add_option("--error-correct-cell",
dest="error_correct_cell",
action="store_true",
help=("Correct errors in the cell barcode"))
group.add_option("--whitelist",
dest="whitelist", type="string",
help=("A whitelist of accepted cell barcodes"))
group.add_option("--blacklist",
dest="blacklist", type="string",
help=("A blacklist of rejected cell barcodes"))
group.add_option("--filter-umi",
dest="filter_umi",
action="store_true",
#help="Filter the UMIs"
help=optparse.SUPPRESS_HELP)
group.add_option("--umi-whitelist", dest="umi_whitelist",
type="string", default=None,
#help="A whitelist of accepted UMIs [default=%default]"
help=optparse.SUPPRESS_HELP)
group.add_option("--umi-whitelist-paired", dest="umi_whitelist_paired",
type="string", default=None,
#help="A whitelist of accepted UMIs for read2[default=%default]"
help=optparse.SUPPRESS_HELP)
group.add_option("--correct-umi-threshold", dest="correct_umi_threshold",
type="int", default=0,
#help="Correct errors in UMIs to the whitelist(s) provided"
#"if within threshold [default=%default]"
help=optparse.SUPPRESS_HELP)
group.add_option("--umi-correct-log", dest="umi_correct_log",
type="string", default=None,
#help="File logging UMI error correction",
help=optparse.SUPPRESS_HELP)
group.add_option("--subset-reads", "--reads-subset",
dest="reads_subset", type="int",
help=("Only extract from the first N reads. If N is "
"greater than the number of reads, all reads will "
"be used"))
group.add_option("--reconcile-pairs",
dest="reconcile", action="store_true",
help=("Allow the presences of reads in read2 input that "
"are not present in read1 input. This allows cell "
"barcode filtering of read1s without "
"considering read2s"))
parser.add_option_group(group)
group = U.OptionGroup(parser, "[EXPERIMENTAl] barcode extraction options")
group.add_option("--either-read", dest="either_read", action="store_true",
help="UMI may be on either read (see "
"--either-read-resolve) for options to resolve cases where"
"UMI is on both reads")
group.add_option("--either-read-resolve",
dest="either_read_resolve", type="choice",
choices=["discard", "quality"],
help=("How to resolve instances where both reads "
"contain a UMI but using --either-read."
"Choose from 'discard' or 'quality'"
"(use highest quality). default=dicard"))
parser.add_option_group(group)
parser.set_defaults(extract_method="string",
filter_cell_barcodes=False,
whitelist=None,
blacklist=None,
error_correct_cell=False,
pattern=None,
pattern2=None,
read2_in=None,
read2_out=False,
read2_stdout=False,
quality_filter_threshold=None,
quality_encoding=None,
reconcile=False,
either_read=False,
either_read_resolve="discard",
ignore_suffix=False)
# add common options (-h/--help, ...) and parse command line
(options, args) = U.Start(parser, argv=argv,
add_extract_options=True,
add_group_dedup_options=False,
add_umi_grouping_options=False,
add_sam_options=False)
if options.filter_cell_barcode:
U.info('Use of --whitelist ensures cell barcodes are filtered. '
'--filter-cell-barcode is no longer required and may be '
'removed in future versions.')
if options.whitelist is not None:
options.filter_cell_barcode = True
if options.retain_umi and not options.extract_method == "regex":
U.error("option --retain-umi only works with --extract-method=regex")
if (options.filtered_out and not options.extract_method == "regex" and
whitelist is None):
U.error("Reads will not be filtered unless extract method is"
"set to regex (--extract-method=regex) or cell"
"barcodes are filtered (--whitelist)")
if options.quality_filter_threshold or options.quality_filter_mask:
if not options.quality_encoding:
U.error("must provide a quality encoding (--quality-"
"encoding) to filter UMIs by quality (--quality"
"-filter-threshold) or mask low quality bases "
"with (--quality-filter-mask)")
extract_cell, extract_umi = U.validateExtractOptions(options)
if options.either_read:
if extract_cell:
U.error("Option to extract from either read (--either-read) "
"is not currently compatible with cell barcode extraction")
if not options.extract_method == "regex":
U.error("Option to extract from either read (--either-read)"
"requires --extract-method=regex")
if not options.pattern or not options.pattern2:
U.error("Option to extract from either read (--either-read)"
"requires --bc-pattern=[PATTERN1] and"
"--bc-pattern2=[PATTERN2]")
if options.filter_umi:
if not options.umi_whitelist:
U.error("must provide a UMI whitelist (--umi-whitelist) if using "
"--filter-umi option")
if options.pattern2 and not options.umi_whitelist_paired:
U.error("must provide a UMI whitelist for paired end "
"(--umi-whitelist-paired) if using --filter-umi option"
"with paired end data")
if not extract_umi:
if options.extract_method == "string":
U.error("barcode pattern(s) do not include any umi bases "
"(marked with 'Ns') %s, %s" % (
options.pattern, options.pattern2))
elif options.extract_method == "regex":
U.error("barcode regex(es) do not include any umi groups "
"(starting with 'umi_') %s, %s" (
options.pattern, options.pattern2))
if options.whitelist:
if not extract_cell:
if options.extract_method == "string":
U.error("barcode pattern(s) do not include any cell bases "
"(marked with 'Cs') %s, %s" % (
options.pattern, options.pattern2))
elif options.extract_method == "regex":
U.error("barcode regex(es) do not include any cell groups "
"(starting with 'cell_') %s, %s" (
options.pattern, options.pattern2))
read1s = umi_methods.fastqIterate(options.stdin)
# set up read extractor
ReadExtractor = extract_methods.ExtractFilterAndUpdate(
options.extract_method,
options.pattern,
options.pattern2,
options.prime3,
extract_cell,
options.quality_encoding,
options.quality_filter_threshold,
options.quality_filter_mask,
options.filter_umi,
options.filter_cell_barcode,
options.retain_umi,
options.either_read,
options.either_read_resolve)
if options.filter_umi:
umi_whitelist, false_to_true_map = whitelist_methods.getUserDefinedBarcodes(
options.umi_whitelist,
options.umi_whitelist_paired,
deriveErrorCorrection=True,
threshold=options.correct_umi_threshold)
U.info("Length of whitelist: %i" % len(umi_whitelist))
U.info("Length of 'correctable' whitelist: %i" % len(false_to_true_map))
ReadExtractor.umi_whitelist = umi_whitelist
ReadExtractor.umi_false_to_true_map = false_to_true_map
ReadExtractor.umi_whitelist_counts = collections.defaultdict(
lambda: collections.Counter())
if options.whitelist:
cell_whitelist, false_to_true_map = whitelist_methods.getUserDefinedBarcodes(
options.whitelist,
getErrorCorrection=options.error_correct_cell)
ReadExtractor.cell_whitelist = cell_whitelist
ReadExtractor.false_to_true_map = false_to_true_map
if options.blacklist:
blacklist = set()
with U.openFile(options.blacklist, "r") as inf:
for line in inf:
blacklist.add(line.strip().split("\t")[0])
ReadExtractor.cell_blacklist = blacklist
# variables for progress monitor
progCount = 0
displayMax = 100000
U.info("Starting barcode extraction")
if options.filtered_out:
filtered_out = U.openFile(options.filtered_out, "w")
if options.read2_in is None:
for read in read1s:
# incrementing count for monitoring progress
progCount += 1
# Update display in every 100kth iteration
if progCount % displayMax == 0:
U.info("Parsed {} reads".format(progCount))
new_read = ReadExtractor(read)
if options.reads_subset:
if (ReadExtractor.read_counts['Input Reads'] >
options.reads_subset):
break
if not new_read:
if options.filtered_out:
filtered_out.write(str(read) + "\n")
continue
options.stdout.write(str(new_read) + "\n")
else:
if options.filtered_out2:
filtered_out2 = U.openFile(options.filtered_out2, "w")
read2s = umi_methods.fastqIterate(U.openFile(options.read2_in))
if options.read2_out:
read2_out = U.openFile(options.read2_out, "w")
if options.reconcile:
strict = False
else:
strict = True
for read1, read2 in umi_methods.joinedFastqIterate(
read1s, read2s, strict, options.ignore_suffix):
# incrementing count for monitoring progress
progCount += 1
# Update display in every 100kth iteration
if progCount % displayMax == 0:
U.info("Parsed {} reads".format(progCount))
sys.stdout.flush()
reads = ReadExtractor(read1, read2)
if options.reads_subset:
if (ReadExtractor.read_counts['Input Reads'] >
options.reads_subset):
break
if not reads:
if options.filtered_out:
filtered_out.write(str(read1) + "\n")
if options.filtered_out2:
filtered_out2.write(str(read2) + "\n")
continue
else:
new_read1, new_read2 = reads
if options.read2_stdout:
options.stdout.write(str(new_read2) + "\n")
else:
options.stdout.write(str(new_read1) + "\n")
if options.read2_out:
read2_out.write(str(new_read2) + "\n")
if options.read2_out:
read2_out.close()
if options.filtered_out:
filtered_out.close()
if options.filtered_out2:
filtered_out2.close()
for k, v in ReadExtractor.getReadCounts().most_common():
U.info("%s: %s" % (k, v))
if options.umi_correct_log:
with U.openFile(options.umi_correct_log, "w") as outf:
outf.write("umi\tcount_no_errors\tcount_errors\n")
for umi, counts in ReadExtractor.umi_whitelist_counts.items():
outf.write("%s\t%i\t%i\n" % (
umi, counts["no_error"], counts["error"]))
outf.close()
U.Stop()
