def align_bam(in_bam, ref_file, names, align_dir, data):
"""Perform realignment of input BAM file; uses unix pipes for avoid IO.
"""
config = data["config"]
out_file = os.path.join(align_dir, "{0}-sort.bam".format(names["lane"]))
novoalign = config_utils.get_program("novoalign", config)
samtools = config_utils.get_program("samtools", config)
resources = config_utils.get_resources("novoalign", config)
num_cores = config["algorithm"].get("num_cores", 1)
max_mem = resources.get("memory", "4G").upper()
extra_novo_args = " ".join(_novoalign_args_from_config(config, False))
if not file_exists(out_file):
with utils.curdir_tmpdir(data, base_dir=align_dir) as work_dir:
with postalign.tobam_cl(data, out_file, bam.is_paired(in_bam)) as (tobam_cl, tx_out_file):
rg_info = get_rg_info(names)
tx_out_prefix = os.path.splitext(tx_out_file)[0]
prefix1 = "%s-in1" % tx_out_prefix
cmd = ("{samtools} sort -n -o -l 0 -@ {num_cores} -m {max_mem} {in_bam} {prefix1} "
"| {novoalign} -o SAM '{rg_info}' -d {ref_file} -f /dev/stdin "
" -F BAMPE -c {num_cores} {extra_novo_args} | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "Novoalign: %s" % names["sample"], None,
[do.file_nonempty(tx_out_file), do.file_reasonable_size(tx_out_file, in_bam)])
return out_file
def align(fastq_file, pair_file, index_dir, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted, deduplicated BAM.
"""
umi_ext = "-cumi" if "umi_bam" in data else ""
out_file = os.path.join(align_dir, "{0}-sort{1}.bam".format(dd.get_sample_name(data), umi_ext))
num_cores = data["config"]["algorithm"].get("num_cores", 1)
rg_info = novoalign.get_rg_info(names)
preset = "sr"
pair_file = pair_file if pair_file else ""
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(fastq_file, pair_file, data)
else:
final_file = None
if not utils.file_exists(out_file) and (final_file is None or not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file != "") as (tobam_cl, tx_out_file):
index_file = None
# Skip trying to use indices now as they provide only slight speed-ups
# and give inconsitent outputs in BAM headers
# If a single index present, index_dir points to that
# if index_dir and os.path.isfile(index_dir):
# index_dir = os.path.dirname(index_dir)
# index_file = os.path.join(index_dir, "%s-%s.mmi" % (dd.get_genome_build(data), preset))
if not index_file or not os.path.exists(index_file):
index_file = dd.get_ref_file(data)
cmd = ("minimap2 -a -x {preset} -R '{rg_info}' -t {num_cores} {index_file} "
"{fastq_file} {pair_file} | ")
do.run(cmd.format(**locals()) + tobam_cl, "minimap2 alignment: %s" % dd.get_sample_name(data))
data["work_bam"] = out_file
return data
def align(fastq_file, pair_file, ref_file, names, align_dir, data,
extra_args=None):
"""Alignment with bowtie2.
"""
config = data["config"]
analysis_config = ANALYSIS.get(data["analysis"].lower())
assert analysis_config, "Analysis %s is not supported by bowtie2" % (data["analysis"])
out_file = os.path.join(align_dir, "{0}-sort.bam".format(dd.get_sample_name(data)))
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(fastq_file, pair_file, data)
else:
final_file = None
if not utils.file_exists(out_file) and (final_file is None or not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file is not None) as (tobam_cl, tx_out_file):
cl = [config_utils.get_program("bowtie2", config)]
cl += _bowtie2_args_from_config(config)
cl += extra_args if extra_args is not None else []
cl += ["-q",
"-x", ref_file]
cl += analysis_config.get("params", [])
if pair_file:
cl += ["-1", fastq_file, "-2", pair_file]
else:
cl += ["-U", fastq_file]
if names and "rg" in names:
cl += ["--rg-id", names["rg"]]
for key, tag in [("sample", "SM"), ("pl", "PL"), ("pu", "PU"), ("lb", "LB")]:
if names.get(key):
cl += ["--rg", "%s:%s" % (tag, names[key])]
cl = [str(i) for i in cl]
cmd = " ".join(cl) + " | " + tobam_cl
do.run(cmd, "Aligning %s and %s with Bowtie2." % (fastq_file, pair_file))
return out_file
def align_pipe(fastq_file, pair_file, ref_file, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted output BAM.
"""
pair_file = pair_file if pair_file else ""
out_file = os.path.join(align_dir, "{0}-sort.bam".format(names["lane"]))
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file = alignprep.split_namedpipe_cl(fastq_file, data)
if pair_file:
pair_file = alignprep.split_namedpipe_cl(pair_file, data)
else:
final_file = None
samtools = config_utils.get_program("samtools", data["config"])
novoalign = config_utils.get_program("novoalign", data["config"])
resources = config_utils.get_resources("novoalign", data["config"])
num_cores = data["config"]["algorithm"].get("num_cores", 1)
max_mem = resources.get("memory", "1G")
extra_novo_args = " ".join(_novoalign_args_from_config(data["config"]))
rg_info = get_rg_info(names)
if not utils.file_exists(out_file) and (final_file is None or not utils.file_exists(final_file)):
with utils.curdir_tmpdir(data) as work_dir:
with postalign.tobam_cl(data, out_file, pair_file != "") as (tobam_cl, tx_out_file):
tx_out_prefix = os.path.splitext(tx_out_file)[0]
cmd = ("{novoalign} -o SAM '{rg_info}' -d {ref_file} -f {fastq_file} {pair_file} "
" -c {num_cores} {extra_novo_args} | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "Novoalign: %s" % names["sample"], None,
[do.file_nonempty(tx_out_file), do.file_reasonable_size(tx_out_file, fastq_file)])
data["work_bam"] = out_file
return data
def align_bam(in_bam, ref_file, names, align_dir, data):
"""Perform direct alignment of an input BAM file with BWA using pipes.
This avoids disk IO by piping between processes:
- samtools sort of input BAM to queryname
- bedtools conversion to interleaved FASTQ
- bwa-mem alignment
- samtools conversion to BAM
- samtools sort to coordinate
"""
config = data["config"]
out_file = os.path.join(align_dir, "{0}-sort.bam".format(names["lane"]))
samtools = config_utils.get_program("samtools", config)
bedtools = config_utils.get_program("bedtools", config)
bwa = config_utils.get_program("bwa", config)
resources = config_utils.get_resources("samtools", config)
num_cores = config["algorithm"].get("num_cores", 1)
# adjust memory for samtools since used for input and output
max_mem = config_utils.adjust_memory(resources.get("memory", "1G"),
3, "decrease")
rg_info = novoalign.get_rg_info(names)
if not utils.file_exists(out_file):
with utils.curdir_tmpdir() as work_dir:
with postalign.tobam_cl(data, out_file, bam.is_paired(in_bam)) as (tobam_cl, tx_out_file):
tx_out_prefix = os.path.splitext(tx_out_file)[0]
prefix1 = "%s-in1" % tx_out_prefix
cmd = ("{samtools} sort -n -o -l 0 -@ {num_cores} -m {max_mem} {in_bam} {prefix1} "
"| {bedtools} bamtofastq -i /dev/stdin -fq /dev/stdout -fq2 /dev/stdout "
"| {bwa} mem -p -M -t {num_cores} -R '{rg_info}' -v 1 {ref_file} - | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa mem alignment from BAM: %s" % names["sample"], None,
[do.file_nonempty(tx_out_file), do.file_reasonable_size(tx_out_file, in_bam)])
return out_file
def _align_mem(fastq_file, pair_file, ref_file, out_file, names, rg_info, data):
"""Perform bwa-mem alignment on supported read lengths.
"""
with postalign.tobam_cl(data, out_file, pair_file != "") as (tobam_cl, tx_out_file):
cmd = "%s | %s" % (_get_bwa_mem_cmd(data, out_file, ref_file, fastq_file, pair_file), tobam_cl)
do.run(cmd, "bwa mem alignment from fastq: %s" % names["sample"], None,
[do.file_nonempty(tx_out_file), do.file_reasonable_size(tx_out_file, fastq_file)])
return out_file
def _align_mem(fastq_file, pair_file, ref_file, out_file, names, rg_info, data):
"""Perform bwa-mem alignment on supported read lengths.
"""
bwa = config_utils.get_program("bwa", data["config"])
num_cores = data["config"]["algorithm"].get("num_cores", 1)
with utils.curdir_tmpdir() as work_dir:
with postalign.tobam_cl(data, out_file, pair_file != "") as (tobam_cl, tx_out_file):
cmd = ("{bwa} mem -M -t {num_cores} -R '{rg_info}' -v 1 {ref_file} "
"{fastq_file} {pair_file} | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa mem alignment from fastq: %s" % names["sample"], None,
[do.file_nonempty(tx_out_file), do.file_reasonable_size(tx_out_file, fastq_file)])
return out_file
def align(fastq_file, pair_file, ref_file, names, align_dir, data,
extra_args=None):
"""Do standard or paired end alignment with bowtie.
"""
num_hits = 1
if data["analysis"].lower().startswith("smallrna-seq"):
num_hits = 1000
config = data['config']
out_file = os.path.join(align_dir, "{0}-sort.bam".format(dd.get_sample_name(data)))
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(fastq_file, pair_file, data)
else:
final_file = None
if fastq_file.endswith(".gz"):
fastq_file = "<(gunzip -c %s)" % fastq_file
if pair_file:
pair_file = "<(gunzip -c %s)" % pair_file
if not utils.file_exists(out_file) and (final_file is None or not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file is not None) as (tobam_cl, tx_out_file):
cl = [config_utils.get_program("bowtie", config)]
cl += _bowtie_args_from_config(data)
cl += extra_args if extra_args is not None else []
cl += ["-q",
"-v", 2,
"-k", num_hits,
"-X", 2000, # default is too selective for most data
"--best",
"--strata",
"--sam",
ref_file]
if pair_file:
cl += ["-1", fastq_file, "-2", pair_file]
else:
cl += [fastq_file]
cl = [str(i) for i in cl]
fix_rg_cmd = r"samtools addreplacerg -r '%s' -" % novoalign.get_rg_info(data["rgnames"])
if fix_rg_cmd:
cmd = " ".join(cl) + " | " + fix_rg_cmd + " | " + tobam_cl
else:
cmd = " ".join(cl) + " | " + tobam_cl
do.run(cmd, "Running Bowtie on %s and %s." % (fastq_file, pair_file), data)
return out_file
def align(fastq_file, pair_file, index_dir, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted, deduplicated BAM.
Pipes in input, handling paired and split inputs, using interleaving magic
from: https://biowize.wordpress.com/2015/03/26/the-fastest-darn-fastq-decoupling-procedure-i-ever-done-seen/
"""
out_file = os.path.join(align_dir, "{0}-sort.bam".format(dd.get_sample_name(data)))
num_cores = data["config"]["algorithm"].get("num_cores", 1)
resources = config_utils.get_resources("snap", data["config"])
rg_info = novoalign.get_rg_info(names)
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(fastq_file, pair_file, data)
fastq_file = fastq_file[2:-1]
if pair_file:
pair_file = pair_file[2:-1]
stream_input = (r"paste <({fastq_file} | paste - - - -) "
r"<({pair_file} | paste - - - -) | tr '\t' '\n'")
else:
stream_input = fastq_file[2:-1]
else:
assert fastq_file.endswith(".gz")
if pair_file:
stream_input = (r"paste <(zcat {fastq_file} | paste - - - -) "
r"<(zcat {pair_file} | paste - - - -) | tr '\t' '\n'")
else:
stream_input = "zcat {fastq_file}"
pair_file = pair_file if pair_file else ""
if not utils.file_exists(out_file) and (final_file is None or not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file is not None) as (tobam_cl, tx_out_file):
if pair_file:
sub_cmd = "paired"
input_cmd = "-pairedInterleavedFastq -"
else:
sub_cmd = "single"
input_cmd = "-fastq -"
stream_input = stream_input.format(**locals())
cmd = ("{stream_input} | snap-aligner {sub_cmd} {index_dir} {input_cmd} "
"-R '{rg_info}' -t {num_cores} -M -o -sam - | ")
do.run(cmd.format(**locals()) + tobam_cl, "SNAP alignment: %s" % names["sample"])
data["work_bam"] = out_file
return data
def _align_mem(fastq_file, pair_file, ref_file, out_file, names, rg_info,
data):
"""Perform bwa-mem alignment on supported read lengths.
"""
with postalign.tobam_cl(data, out_file,
pair_file != "") as (tobam_cl, tx_out_file):
cmd = ("unset JAVA_HOME && "
"%s | %s" % (_get_bwa_mem_cmd(data,
out_file,
ref_file,
fastq_file,
pair_file,
with_hla=False), tobam_cl))
do.run(cmd, "bwa mem alignment from fastq: %s" % names["sample"], None,
[
do.file_nonempty(tx_out_file),
do.file_reasonable_size(tx_out_file, fastq_file)
])
return out_file
def align(fastq_file, pair_file, ref_file, names, align_dir, data):
paired = True if pair_file else False
hisat2 = config_utils.get_program("hisat2", data)
num_cores = dd.get_num_cores(data)
quality_flag = _get_quality_flag(data)
stranded_flag = _get_stranded_flag(data, paired)
rg_flags = _get_rg_flags(names)
out_file = os.path.join(align_dir, "{0}-sort.bam".format(dd.get_sample_name(data)))
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(fastq_file, pair_file, data)
else:
final_file = None
if not file_exists(out_file) and (final_file is None or not file_exists(final_file)):
cmd = ("{hisat2} --new-summary -x {ref_file} -p {num_cores} {quality_flag} {stranded_flag} "
"{rg_flags} ")
if paired:
cmd += "-1 {fastq_file} -2 {pair_file} "
else:
cmd += "-U {fastq_file} "
if dd.get_analysis(data).lower() == "smallrna-seq":
cmd += "-k 1000 "
# if assembling transcripts, set flags that cufflinks/stringtie can use
if dd.get_transcript_assembler(data):
cmd += "--dta-cufflinks "
if dd.get_analysis(data).lower() == "rna-seq":
splicesites = get_known_splicesites_file(align_dir, data)
if file_exists(splicesites):
cmd += "--known-splicesite-infile {splicesites} "
novel_splicesite_file = os.path.join(align_dir, "{0}-novelsplicesites.bed".format(dd.get_sample_name(data)))
cmd += "--novel-splicesite-outfile {novel_splicesite_file} "
# apply additional hisat2 options
cmd += " ".join(_get_options_from_config(data))
message = "Aligning %s and %s with hisat2." % (fastq_file, pair_file)
with postalign.tobam_cl(data, out_file, pair_file is not None) as (tobam_cl, tx_out_file):
cmd += " | " + tobam_cl
do.run(cmd.format(**locals()), message)
data = dd.set_work_bam(data, out_file)
junctionbed = get_splicejunction_file(align_dir, data)
data = dd.set_junction_bed(data, junctionbed)
return data
def _align_mem(fastq_file, pair_file, ref_file, out_file, names, rg_info,
data):
"""Perform bwa-mem alignment on supported read lengths.
"""
bwa = config_utils.get_program("bwa", data["config"])
num_cores = data["config"]["algorithm"].get("num_cores", 1)
with utils.curdir_tmpdir(data) as work_dir:
with postalign.tobam_cl(data, out_file,
pair_file != "") as (tobam_cl, tx_out_file):
cmd = (
"{bwa} mem -M -t {num_cores} -R '{rg_info}' -v 1 {ref_file} "
"{fastq_file} {pair_file} | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa mem alignment from fastq: %s" % names["sample"],
None, [
do.file_nonempty(tx_out_file),
do.file_reasonable_size(tx_out_file, fastq_file)
])
return out_file
def _align_backtrack(fastq_file, pair_file, ref_file, out_file, names, rg_info, data):
"""Perform a BWA alignment using 'aln' backtrack algorithm.
"""
bwa = config_utils.get_program("bwa", data["config"])
config = data["config"]
sai1_file = "%s_1.sai" % os.path.splitext(out_file)[0]
sai2_file = "%s_2.sai" % os.path.splitext(out_file)[0] if pair_file else ""
if not utils.file_exists(sai1_file):
with file_transaction(data, sai1_file) as tx_sai1_file:
_run_bwa_align(fastq_file, ref_file, tx_sai1_file, config)
if sai2_file and not utils.file_exists(sai2_file):
with file_transaction(data, sai2_file) as tx_sai2_file:
_run_bwa_align(pair_file, ref_file, tx_sai2_file, config)
with postalign.tobam_cl(data, out_file, pair_file != "") as (tobam_cl, tx_out_file):
align_type = "sampe" if sai2_file else "samse"
cmd = ("{bwa} {align_type} -r '{rg_info}' {ref_file} {sai1_file} {sai2_file} "
"{fastq_file} {pair_file} | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa %s" % align_type, data)
return out_file
def _align_backtrack(fastq_file, pair_file, ref_file, out_file, names, rg_info, data):
"""Perform a BWA alignment using 'aln' backtrack algorithm.
"""
assert not data.get("align_split"), "Do not handle split alignments with non-piped bwa"
bwa = config_utils.get_program("bwa", data["config"])
config = data["config"]
sai1_file = "%s_1.sai" % os.path.splitext(out_file)[0]
sai2_file = "%s_2.sai" % os.path.splitext(out_file)[0] if pair_file else ""
if not utils.file_exists(sai1_file):
with file_transaction(data, sai1_file) as tx_sai1_file:
_run_bwa_align(fastq_file, ref_file, tx_sai1_file, config)
if sai2_file and not utils.file_exists(sai2_file):
with file_transaction(data, sai2_file) as tx_sai2_file:
_run_bwa_align(pair_file, ref_file, tx_sai2_file, config)
with postalign.tobam_cl(data, out_file, pair_file != "") as (tobam_cl, tx_out_file):
align_type = "sampe" if sai2_file else "samse"
cmd = "{bwa} {align_type} -r '{rg_info}' {ref_file} {sai1_file} {sai2_file} " "{fastq_file} {pair_file} | "
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa %s" % align_type, data)
return out_file
def align_pipe(fastq_file, pair_file, ref_file, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted output BAM.
"""
pair_file = pair_file if pair_file else ""
# back compatible -- older files were named with lane information, use sample name now
out_file = os.path.join(align_dir, "{0}-sort.bam".format(names["lane"]))
if not utils.file_exists(out_file):
out_file = os.path.join(
align_dir, "{0}-sort.bam".format(dd.get_sample_name(data)))
if data.get("align_split") or fastq_file.endswith(".sdf"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(
fastq_file, pair_file, data)
else:
final_file = None
samtools = config_utils.get_program("samtools", data["config"])
novoalign = config_utils.get_program("novoalign", data["config"])
resources = config_utils.get_resources("novoalign", data["config"])
num_cores = data["config"]["algorithm"].get("num_cores", 1)
max_mem = resources.get("memory", "1G")
extra_novo_args = " ".join(_novoalign_args_from_config(data["config"]))
rg_info = get_rg_info(names)
if not utils.file_exists(out_file) and (final_file is None or
not utils.file_exists(final_file)):
with tx_tmpdir(data) as work_dir:
with postalign.tobam_cl(data, out_file,
pair_file != "") as (tobam_cl,
tx_out_file):
tx_out_prefix = os.path.splitext(tx_out_file)[0]
cmd = (
"unset JAVA_HOME && "
"{novoalign} -o SAM '{rg_info}' -d {ref_file} -f {fastq_file} {pair_file} "
" -c {num_cores} {extra_novo_args} | ")
cmd = (cmd + tobam_cl).format(**locals())
do.run(cmd, "Novoalign: %s" % names["sample"], None, [
do.file_nonempty(tx_out_file),
do.file_reasonable_size(tx_out_file, fastq_file)
])
data["work_bam"] = out_file
return data
def align(fastq_file, pair_file, index_dir, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted, deduplicated BAM.
"""
umi_ext = "-cumi" if "umi_bam" in data else ""
out_file = os.path.join(
align_dir, "{0}-sort{1}.bam".format(dd.get_sample_name(data), umi_ext))
num_cores = data["config"]["algorithm"].get("num_cores", 1)
rg_info = novoalign.get_rg_info(names)
preset = "sr"
pair_file = pair_file if pair_file else ""
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(
fastq_file, pair_file, data)
else:
final_file = None
if not utils.file_exists(out_file) and (final_file is None or
not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file,
pair_file != "") as (tobam_cl, tx_out_file):
index_file = None
# Skip trying to use indices now as they provide only slight speed-ups
# and give inconsitent outputs in BAM headers
# If a single index present, index_dir points to that
# if index_dir and os.path.isfile(index_dir):
# index_dir = os.path.dirname(index_dir)
# index_file = os.path.join(index_dir, "%s-%s.mmi" % (dd.get_genome_build(data), preset))
if not index_file or not os.path.exists(index_file):
index_file = dd.get_ref_file(data)
cmd = (
"minimap2 -a -x {preset} -R '{rg_info}' -t {num_cores} {index_file} "
"{fastq_file} {pair_file} | ")
do.run(
cmd.format(**locals()) + tobam_cl,
"minimap2 alignment: %s" % dd.get_sample_name(data))
data["work_bam"] = out_file
return data
def align_bam(in_bam, ref_file, names, align_dir, data):
"""Perform direct alignment of an input BAM file with BWA using pipes.
This avoids disk IO by piping between processes:
- samtools sort of input BAM to queryname
- bedtools conversion to interleaved FASTQ
- bwa-mem alignment
- samtools conversion to BAM
- samtools sort to coordinate
"""
config = data["config"]
out_file = os.path.join(align_dir, "{0}-sort.bam".format(names["lane"]))
samtools = config_utils.get_program("samtools", config)
bedtools = config_utils.get_program("bedtools", config)
bwa = config_utils.get_program("bwa", config)
resources = config_utils.get_resources("samtools", config)
num_cores = config["algorithm"].get("num_cores", 1)
# adjust memory for samtools since used for input and output
max_mem = config_utils.adjust_memory(resources.get("memory", "1G"), 3,
"decrease").upper()
rg_info = novoalign.get_rg_info(names)
if not utils.file_exists(out_file):
with utils.curdir_tmpdir(data) as work_dir:
with postalign.tobam_cl(data, out_file,
bam.is_paired(in_bam)) as (tobam_cl,
tx_out_file):
tx_out_prefix = os.path.splitext(tx_out_file)[0]
prefix1 = "%s-in1" % tx_out_prefix
cmd = (
"{samtools} sort -n -o -l 1 -@ {num_cores} -m {max_mem} {in_bam} {prefix1} "
"| {bedtools} bamtofastq -i /dev/stdin -fq /dev/stdout -fq2 /dev/stdout "
"| {bwa} mem -p -M -t {num_cores} -R '{rg_info}' -v 1 {ref_file} - | "
)
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa mem alignment from BAM: %s" % names["sample"],
None, [
do.file_nonempty(tx_out_file),
do.file_reasonable_size(tx_out_file, in_bam)
])
return out_file
def align(fastq_file, pair_file, index_dir, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted, deduplicated BAM.
TODO: Use streaming with new development version of SNAP to feed into
structural variation preparation de-duplication.
"""
pair_file = pair_file if pair_file else ""
out_file = os.path.join(align_dir, "{0}-sort.bam".format(names["lane"]))
assert not data.get("align_split"), "Split alignments not supported with SNAP"
snap = config_utils.get_program("snap", data["config"])
num_cores = data["config"]["algorithm"].get("num_cores", 1)
resources = config_utils.get_resources("snap", data["config"])
rg_info = novoalign.get_rg_info(names)
is_paired = bam.is_paired(fastq_file) if fastq_file.endswith(".bam") else pair_file
if not utils.file_exists(out_file):
with postalign.tobam_cl(data, out_file, is_paired) as (tobam_cl, tx_out_file):
cmd_name = "paired" if is_paired else "single"
cmd = ("{snap} {cmd_name} {index_dir} {fastq_file} {pair_file} "
"-R '{rg_info}' -t {num_cores} -M -o -sam - | ")
do.run(cmd.format(**locals()) + tobam_cl, "SNAP alignment: %s" % names["sample"])
data["work_bam"] = out_file
return data
def align(fastq_file, pair_file, index_dir, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted, deduplicated BAM.
"""
umi_ext = "-cumi" if "umi_bam" in data else ""
out_file = os.path.join(align_dir, "{0}-sort{1}.bam".format(dd.get_sample_name(data), umi_ext))
num_cores = data["config"]["algorithm"].get("num_cores", 1)
rg_info = "rgid={rg} rgpl={pl} rgpu={pu} rgsm={sample}".format(**names)
pair_file = pair_file if pair_file else ""
final_file = None
if data.get("align_split"):
# BBMap does not accept input fastq streams
raise ValueError("bbmap is not compatible with alignment splitting, set `align_split: false`")
pair_arg = "in2=%s" % pair_file if pair_file else ""
if not utils.file_exists(out_file) and (final_file is None or not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file != "") as (tobam_cl, tx_out_file):
if index_dir.endswith(("/ref", "/ref/")):
index_dir = os.path.dirname(index_dir)
# sam=1.3 required for compatibility with strelka2
cmd = ("bbmap.sh sam=1.3 mdtag=t {rg_info} path={index_dir} in1={fastq_file} "
"{pair_arg} out=stdout.sam | ")
do.run(cmd.format(**locals()) + tobam_cl, "bbmap alignment: %s" % dd.get_sample_name(data))
data["work_bam"] = out_file
return data
def align(fastq_file,
pair_file,
ref_file,
names,
align_dir,
data,
extra_args=None):
"""Do standard or paired end alignment with bowtie.
"""
num_hits = 1
if data["analysis"].lower().startswith("smallrna-seq"):
num_hits = 1000
config = data['config']
out_file = os.path.join(align_dir,
"{0}-sort.bam".format(dd.get_sample_name(data)))
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(
fastq_file, pair_file, data)
else:
final_file = None
if fastq_file.endswith(".gz"):
fastq_file = "<(gunzip -c %s)" % fastq_file
if pair_file:
pair_file = "<(gunzip -c %s)" % pair_file
if not utils.file_exists(out_file) and (final_file is None or
not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file
is not None) as (tobam_cl, tx_out_file):
cl = [config_utils.get_program("bowtie", config)]
cl += _bowtie_args_from_config(data)
cl += extra_args if extra_args is not None else []
cl += [
"-q",
"-v",
2,
"-k",
num_hits,
"-X",
2000, # default is too selective for most data
"--best",
"--strata",
"--sam",
ref_file
]
if pair_file:
cl += ["-1", fastq_file, "-2", pair_file]
else:
cl += [fastq_file]
cl = [str(i) for i in cl]
fix_rg_cmd = r"samtools addreplacerg -r '%s' -" % novoalign.get_rg_info(
data["rgnames"])
if fix_rg_cmd:
cmd = " ".join(cl) + " | " + fix_rg_cmd + " | " + tobam_cl
else:
cmd = " ".join(cl) + " | " + tobam_cl
do.run(cmd,
"Running Bowtie on %s and %s." % (fastq_file, pair_file),
data)
return out_file
def align_bam(in_bam, ref_file, names, align_dir, data):
"""Perform direct alignment of an input BAM file with BWA using pipes.
This avoids disk IO by piping between processes:
- samtools sort of input BAM to queryname
- bedtools conversion to interleaved FASTQ
- bwa-mem alignment
- samtools conversion to BAM
- samtools sort to coordinate
"""
config = data["config"]
out_file = os.path.join(align_dir, "{0}-sort.bam".format(names["lane"]))
samtools = config_utils.get_program("samtools", config)
bedtools = config_utils.get_program("bedtools", config)
resources = config_utils.get_resources("samtools", config)
num_cores = config["algorithm"].get("num_cores", 1)
# adjust memory for samtools since used for input and output
max_mem = config_utils.adjust_memory(resources.get("memory", "1G"), 3,
"decrease").upper()
if not utils.file_exists(out_file):
with tx_tmpdir(data) as work_dir:
with postalign.tobam_cl(data, out_file,
bam.is_paired(in_bam)) as (tobam_cl,
tx_out_file):
if not hla_on(data) or needs_separate_hla(data):
bwa_cmd = _get_bwa_mem_cmd(data,
out_file,
ref_file,
"-",
with_hla=False)
else:
bwa_cmd = _get_bwa_mem_cmd(data,
out_file,
ref_file,
"-",
with_hla=True)
tx_out_prefix = os.path.splitext(tx_out_file)[0]
prefix1 = "%s-in1" % tx_out_prefix
cmd = (
"unset JAVA_HOME && "
"{samtools} sort -n -l 1 -@ {num_cores} -m {max_mem} {in_bam} -T {prefix1} "
"| {bedtools} bamtofastq -i /dev/stdin -fq /dev/stdout -fq2 /dev/stdout "
"| {bwa_cmd} | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa mem alignment from BAM: %s" % names["sample"],
None, [
do.file_nonempty(tx_out_file),
do.file_reasonable_size(tx_out_file, in_bam)
])
data["work_bam"] = out_file
hla_file = "HLA-" + out_file
if needs_separate_hla(data) and not utils.file_exists(hla_file):
with tx_tmpdir(data) as work_dir:
with postalign.tobam_cl(data, hla_file,
bam.is_paired(in_bam)) as (tobam_cl,
tx_out_file):
bwa_cmd = _get_bwa_mem_cmd(data,
hla_file,
ref_file,
"-",
with_hla=True)
tx_out_prefix = os.path.splitext(tx_out_file)[0]
prefix1 = "%s-in1" % tx_out_prefix
cmd = (
"unset JAVA_HOME && "
"{samtools} sort -n -l 1 -@ {num_cores} -m {max_mem} {in_bam} -T {prefix1} "
"| {bedtools} bamtofastq -i /dev/stdin -fq /dev/stdout -fq2 /dev/stdout "
"| {bwa_cmd} | ")
cmd = cmd.format(**locals()) + tobam_cl
do.run(cmd, "bwa mem alignment from BAM: %s" % names["sample"],
None, [
do.file_nonempty(tx_out_file),
do.file_reasonable_size(tx_out_file, in_bam)
])
hla_file = _align_mem_hla(fastq_file, pair_file, ref_file, hla_file,
names, rg_info, data)
data["hla_bam"] = hla_file
return data
def align(fastq_file, pair_file, index_dir, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted, deduplicated BAM.
Pipes in input, handling paired and split inputs, using interleaving magic
from: https://biowize.wordpress.com/2015/03/26/the-fastest-darn-fastq-decoupling-procedure-i-ever-done-seen/
Then converts a tab delimited set of outputs into interleaved fastq.
awk changes spaces to underscores since SNAP only takes the initial name.
SNAP requires /1 and /2 at the end of read names. If these are not present
in the initial fastq may need to expand awk code to do this.
"""
out_file = os.path.join(align_dir, "{0}-sort.bam".format(dd.get_sample_name(data)))
num_cores = data["config"]["algorithm"].get("num_cores", 1)
resources = config_utils.get_resources("snap", data["config"])
rg_info = novoalign.get_rg_info(names)
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(fastq_file, pair_file, data)
fastq_file = fastq_file[2:-1]
if pair_file:
pair_file = pair_file[2:-1]
stream_input = (r"paste <({fastq_file} | paste - - - -) "
r"<({pair_file} | paste - - - -) | "
r"""awk 'BEGIN {{FS="\t"; OFS="\n"}} """
r"""{{ """
r"""split($1, P1, " "); split($5, P5, " "); """
r"""if ($1 !~ /\/1$/) $1 = P1[1]"/1"; if ($5 !~ /\/2$/) $5 = P5[1]"/2"; """
r"""gsub(" ", "_", $1); gsub(" ", "_", $5); """
r"""print $1, $2, "+", $4, $5, $6, "+", $8}}' | sponge """)
else:
stream_input = fastq_file[2:-1]
else:
final_file = None
assert fastq_file.endswith(".gz")
if pair_file:
stream_input = (r"paste <(zcat {fastq_file} | paste - - - -) "
r"<(zcat {pair_file} | paste - - - -) | "
r"""awk 'BEGIN {{FS="\t"; OFS="\n"}} """
r"""{{ """
r"""split($1, P1, " "); split($5, P5, " "); """
r"""if ($1 !~ /\/1$/) $1 = P1[1]"/1"; if ($5 !~ /\/2$/) $5 = P5[1]"/2"; """
r"""gsub(" ", "_", $1); gsub(" ", "_", $5); """
r"""print $1, $2, "+", $4, $5, $6, "+", $8}}' | sponge """)
else:
stream_input = "zcat {fastq_file}"
pair_file = pair_file if pair_file else ""
if not utils.file_exists(out_file) and (final_file is None or not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file is not None) as (tobam_cl, tx_out_file):
if pair_file:
sub_cmd = "paired"
input_cmd = "-pairedInterleavedFastq -"
else:
sub_cmd = "single"
input_cmd = "-fastq -"
stream_input = stream_input.format(**locals())
tmp_dir = os.path.dirname(tx_out_file)
cmd = ("export TMPDIR={tmp_dir} && {stream_input} | snap-aligner {sub_cmd} {index_dir} {input_cmd} "
"-R '{rg_info}' -t {num_cores} -M -o -sam - | ")
do.run(cmd.format(**locals()) + tobam_cl, "SNAP alignment: %s" % names["sample"])
data["work_bam"] = out_file
return data
def align(fastq_file, pair_file, index_dir, names, align_dir, data):
"""Perform piped alignment of fastq input files, generating sorted, deduplicated BAM.
Pipes in input, handling paired and split inputs, using interleaving magic
from: https://biowize.wordpress.com/2015/03/26/the-fastest-darn-fastq-decoupling-procedure-i-ever-done-seen/
Then converts a tab delimited set of outputs into interleaved fastq.
awk changes spaces to underscores since SNAP only takes the initial name.
SNAP requires /1 and /2 at the end of read names. If these are not present
in the initial fastq may need to expand awk code to do this.
"""
out_file = os.path.join(align_dir,
"{0}-sort.bam".format(dd.get_sample_name(data)))
num_cores = data["config"]["algorithm"].get("num_cores", 1)
resources = config_utils.get_resources("snap", data["config"])
rg_info = novoalign.get_rg_info(names)
if data.get("align_split"):
final_file = out_file
out_file, data = alignprep.setup_combine(final_file, data)
fastq_file, pair_file = alignprep.split_namedpipe_cls(
fastq_file, pair_file, data)
fastq_file = fastq_file[2:-1]
if pair_file:
pair_file = pair_file[2:-1]
stream_input = (
r"paste <({fastq_file} | paste - - - -) "
r"<({pair_file} | paste - - - -) | "
r"""awk 'BEGIN {{FS="\t"; OFS="\n"}} """
r"""{{ """
r"""split($1, P1, " "); split($5, P5, " "); """
r"""if ($1 !~ /\/1$/) $1 = P1[1]"/1"; if ($5 !~ /\/2$/) $5 = P5[1]"/2"; """
r"""gsub(" ", "_", $1); gsub(" ", "_", $5); """
r"""print $1, $2, "+", $4, $5, $6, "+", $8}}' """)
else:
stream_input = fastq_file[2:-1]
else:
final_file = None
assert fastq_file.endswith(".gz")
if pair_file:
stream_input = (
r"paste <(zcat {fastq_file} | paste - - - -) "
r"<(zcat {pair_file} | paste - - - -) | "
r"""awk 'BEGIN {{FS="\t"; OFS="\n"}} """
r"""{{ """
r"""split($1, P1, " "); split($5, P5, " "); """
r"""if ($1 !~ /\/1$/) $1 = P1[1]"/1"; if ($5 !~ /\/2$/) $5 = P5[1]"/2"; """
r"""gsub(" ", "_", $1); gsub(" ", "_", $5); """
r"""print $1, $2, "+", $4, $5, $6, "+", $8}}' """)
else:
stream_input = "zcat {fastq_file}"
pair_file = pair_file if pair_file else ""
if not utils.file_exists(out_file) and (final_file is None or
not utils.file_exists(final_file)):
with postalign.tobam_cl(data, out_file, pair_file
is not None) as (tobam_cl, tx_out_file):
if pair_file:
sub_cmd = "paired"
input_cmd = "-pairedInterleavedFastq -"
else:
sub_cmd = "single"
input_cmd = "-fastq -"
stream_input = stream_input.format(**locals())
tmp_dir = os.path.dirname(tx_out_file)
cmd = (
"export TMPDIR={tmp_dir} && unset JAVA_HOME && {stream_input} | "
"snap-aligner {sub_cmd} {index_dir} {input_cmd} "
"-R '{rg_info}' -t {num_cores} -M -o -sam - | ")
do.run(
cmd.format(**locals()) + tobam_cl,
"SNAP alignment: %s" % names["sample"])
data["work_bam"] = out_file
return data
