def gatk_realigner(align_bam, ref_file, config, dbsnp=None, region=None,
out_file=None, deep_coverage=False):
"""Realign a BAM file around indels using GATK, returning sorted BAM.
"""
runner = broad.runner_from_config(config)
bam.index(align_bam, config)
runner.run_fn("picard_index_ref", ref_file)
ref.fasta_idx(ref_file)
if region:
align_bam = subset_bam_by_region(align_bam, region, out_file)
bam.index(align_bam, config)
if has_aligned_reads(align_bam, region):
variant_regions = config["algorithm"].get("variant_regions", None)
realign_target_file = gatk_realigner_targets(runner, align_bam,
ref_file, dbsnp, region,
out_file, deep_coverage,
variant_regions)
realign_bam = gatk_indel_realignment(runner, align_bam, ref_file,
realign_target_file, region,
out_file, deep_coverage)
# No longer required in recent GATK (> Feb 2011) -- now done on the fly
# realign_sort_bam = runner.run_fn("picard_fixmate", realign_bam)
return realign_bam
elif out_file:
shutil.copy(align_bam, out_file)
return out_file
else:
return align_bam
def gatk_realigner(align_bam,
ref_file,
config,
dbsnp=None,
region=None,
out_file=None,
deep_coverage=False):
"""Realign a BAM file around indels using GATK, returning sorted BAM.
"""
runner = broad.runner_from_config(config)
bam.index(align_bam, config)
runner.run_fn("picard_index_ref", ref_file)
ref.fasta_idx(ref_file)
if region:
align_bam = subset_bam_by_region(align_bam, region, out_file)
bam.index(align_bam, config)
if has_aligned_reads(align_bam, region):
variant_regions = config["algorithm"].get("variant_regions", None)
realign_target_file = gatk_realigner_targets(runner, align_bam,
ref_file, dbsnp, region,
out_file, deep_coverage,
variant_regions)
realign_bam = gatk_indel_realignment(runner, align_bam, ref_file,
realign_target_file, region,
out_file, deep_coverage)
# No longer required in recent GATK (> Feb 2011) -- now done on the fly
# realign_sort_bam = runner.run_fn("picard_fixmate", realign_bam)
return realign_bam
elif out_file:
shutil.copy(align_bam, out_file)
return out_file
else:
return align_bam
def _get_coverage_file(in_bam, ref_file, region, region_file, depth, base_file,
data):
"""Retrieve summary of coverage in a region.
Requires positive non-zero mapping quality at a position, matching GATK's
CallableLoci defaults.
"""
out_file = "%s-genomecov.bed" % utils.splitext_plus(base_file)[0]
if not utils.file_exists(out_file):
with file_transaction(data, out_file) as tx_out_file:
fai_file = ref.fasta_idx(ref_file, data["config"])
sambamba = config_utils.get_program("sambamba", data["config"])
bedtools = config_utils.get_program("bedtools", data["config"])
cmd = (
"{sambamba} view -F 'mapping_quality > 0' -L {region_file} -f bam -l 1 {in_bam} | "
"{bedtools} genomecov -split -ibam stdin -bga -g {fai_file} "
"> {tx_out_file}")
do.run(cmd.format(**locals()),
"bedtools genomecov: %s" % (str(region)), data)
# Empty output file, no coverage for the whole contig
if not utils.file_exists(out_file):
with file_transaction(data, out_file) as tx_out_file:
with open(tx_out_file, "w") as out_handle:
for feat in get_ref_bedtool(ref_file, data["config"], region):
out_handle.write("%s\t%s\t%s\t%s\n" %
(feat.chrom, feat.start, feat.end, 0))
return out_file
def add_genes(in_file, data, max_distance=10000):
"""Add gene annotations to a BED file from pre-prepared RNA-seq data.
max_distance -- only keep annotations within this distance of event
"""
gene_file = regions.get_sv_bed(data, "exons", out_dir=os.path.dirname(in_file))
if gene_file and utils.file_exists(in_file):
out_file = "%s-annotated.bed" % utils.splitext_plus(in_file)[0]
if not utils.file_uptodate(out_file, in_file):
input_rec = iter(pybedtools.BedTool(in_file)).next()
# keep everything after standard chrom/start/end, 1-based
extra_fields = range(4, len(input_rec.fields) + 1)
# keep the new gene annotation
gene_index = len(input_rec.fields) + 4
extra_fields.append(gene_index)
columns = ",".join([str(x) for x in extra_fields])
max_column = max(extra_fields) + 1
ops = ",".join(["distinct"] * len(extra_fields))
fai_file = ref.fasta_idx(dd.get_ref_file(data))
with file_transaction(data, out_file) as tx_out_file:
# swap over gene name to '.' if beyond maximum distance
# cut removes the last distance column which can cause issues
# with bedtools merge: 'ERROR: illegal character '.' found in integer conversion of string'
distance_filter = (r"""awk -F$'\t' -v OFS='\t' '{if ($NF > %s) $%s = "."} {print}'""" %
(max_distance, gene_index))
sort_cmd = bedutils.get_sort_cmd()
cmd = ("{sort_cmd} -k1,1 -k2,2n {in_file} | "
"bedtools closest -g <(cut -f1,2 {fai_file} | {sort_cmd} -k1,1 -k2,2n) "
"-d -t all -a - -b <({sort_cmd} -k1,1 -k2,2n {gene_file}) | "
"{distance_filter} | cut -f 1-{max_column} | "
"bedtools merge -i - -c {columns} -o {ops} -delim ',' > {tx_out_file}")
do.run(cmd.format(**locals()), "Annotate BED file with gene info")
return out_file
else:
return in_file
def ref_file_from_bam(bam_file, data):
"""Subset a fasta input file to only a fraction of input contigs.
"""
new_ref = os.path.join(utils.safe_makedir(os.path.join(dd.get_work_dir(data), "inputs", "ref")),
"%s-subset.fa" % dd.get_genome_build(data))
if not utils.file_exists(new_ref):
with file_transaction(data, new_ref) as tx_out_file:
contig_file = "%s-contigs.txt" % utils.splitext_plus(new_ref)[0]
with open(contig_file, "w") as out_handle:
for contig in [x.contig for x in idxstats(bam_file, data) if x.contig != "*"]:
out_handle.write("%s\n" % contig)
cmd = "seqtk subseq -l 100 %s %s > %s" % (dd.get_ref_file(data), contig_file, tx_out_file)
do.run(cmd, "Subset %s to BAM file contigs" % dd.get_genome_build(data))
ref.fasta_idx(new_ref, data["config"])
runner = broad.runner_from_path("picard", data["config"])
runner.run_fn("picard_index_ref", new_ref)
return {"base": new_ref}
def ref_file_from_bam(bam_file, data):
"""Subset a fasta input file to only a fraction of input contigs.
"""
new_ref = os.path.join(utils.safe_makedir(os.path.join(dd.get_work_dir(data), "inputs", "ref")),
"%s-subset.fa" % dd.get_genome_build(data))
if not utils.file_exists(new_ref):
with file_transaction(data, new_ref) as tx_out_file:
contig_file = "%s-contigs.txt" % utils.splitext_plus(new_ref)[0]
with open(contig_file, "w") as out_handle:
for contig in [x.contig for x in idxstats(bam_file, data) if x.contig != "*"]:
out_handle.write("%s\n" % contig)
cmd = "seqtk subseq -l 100 %s %s > %s" % (dd.get_ref_file(data), contig_file, tx_out_file)
do.run(cmd, "Subset %s to BAM file contigs" % dd.get_genome_build(data))
ref.fasta_idx(new_ref, data["config"])
runner = broad.runner_from_path("picard", data["config"])
runner.run_fn("picard_index_ref", new_ref)
return {"base": new_ref}
def get_padded_bed_file(out_dir, bed_file, padding, data):
out_file = os.path.join(out_dir, "%s-padded.bed" % (utils.splitext_plus(os.path.basename(bed_file))[0]))
if utils.file_uptodate(out_file, bed_file):
return out_file
fai_file = ref.fasta_idx(dd.get_ref_file(data))
with file_transaction(data, out_file) as tx_out_file:
cmd = "bedtools slop -i {bed_file} -g {fai_file} -b {padding} > {tx_out_file}"
do.run(cmd.format(**locals()), "Pad BED file", data)
return out_file
def get_padded_bed_file(out_dir, bed_file, padding, data):
bedtools = config_utils.get_program("bedtools", data, default="bedtools")
out_file = os.path.join(out_dir, "%s-padded.bed" % (utils.splitext_plus(os.path.basename(bed_file))[0]))
if utils.file_uptodate(out_file, bed_file):
return out_file
fai_file = ref.fasta_idx(dd.get_ref_file(data))
with file_transaction(data, out_file) as tx_out_file:
cmd = "{bedtools} slop -i {bed_file} -g {fai_file} -b {padding} | bedtools merge -i - > {tx_out_file}"
do.run(cmd.format(**locals()), "Pad BED file", data)
return out_file
def _subset_bed_by_region(in_file, out_file, regions, ref_file, do_merge=True):
orig_bed = pybedtools.BedTool(in_file)
region_bed = pybedtools.BedTool("\n".join(["%s\t%s\t%s" % (c, s, e) for c, s, e in regions]) + "\n",
from_string=True)
sort_kwargs = {"faidx": ref.fasta_idx(ref_file)} if ref_file else {}
if do_merge:
orig_bed.intersect(region_bed, nonamecheck=True).saveas().sort(**sort_kwargs).saveas().\
filter(lambda x: len(x) > 1).saveas().merge().saveas(out_file)
else:
orig_bed.intersect(region_bed, nonamecheck=True).saveas().sort(**sort_kwargs).saveas().\
filter(lambda x: len(x) > 1).saveas(out_file)
def get_padded_bed_file(bed_file, padding, data, bedprep_dir=None):
if not bedprep_dir:
bedprep_dir = utils.safe_makedir(os.path.join(data["dirs"]["work"], "bedprep"))
out_file = os.path.join(bedprep_dir, "%s-padded.bed" % (utils.splitext_plus(os.path.basename(bed_file))[0]))
if utils.file_uptodate(out_file, bed_file):
return out_file
fai_file = ref.fasta_idx(dd.get_ref_file(data))
with file_transaction(data, out_file) as tx_out_file:
cmd = "bedtools slop -i {bed_file} -g {fai_file} -b {padding} > {tx_out_file}"
do.run(cmd.format(**locals()), "Pad BED file", data)
return out_file
def fai_from_bam(ref_file, bam_file, out_file, data):
"""Create a fai index with only contigs in the input BAM file.
"""
contigs = set([x.contig for x in idxstats(bam_file, data)])
if not utils.file_uptodate(out_file, bam_file):
with open(ref.fasta_idx(ref_file, data["config"])) as in_handle:
with file_transaction(data, out_file) as tx_out_file:
with open(tx_out_file, "w") as out_handle:
for line in (l for l in in_handle if l.strip()):
if line.split()[0] in contigs:
out_handle.write(line)
return out_file
def fai_from_bam(ref_file, bam_file, out_file, data):
"""Create a fai index with only contigs in the input BAM file.
"""
contigs = set([x.contig for x in idxstats(bam_file, data)])
if not utils.file_uptodate(out_file, bam_file):
with open(ref.fasta_idx(ref_file, data["config"])) as in_handle:
with file_transaction(data, out_file) as tx_out_file:
with open(tx_out_file, "w") as out_handle:
for line in (l for l in in_handle if l.strip()):
if line.split()[0] in contigs:
out_handle.write(line)
return out_file
def _prep_callable_bed(in_file, work_dir, stats, data):
"""Sort and merge callable BED regions to prevent SV double counting
"""
out_file = os.path.join(work_dir, "%s-merge.bed.gz" % utils.splitext_plus(os.path.basename(in_file))[0])
gsort = config_utils.get_program("gsort", data)
if not utils.file_uptodate(out_file, in_file):
with file_transaction(data, out_file) as tx_out_file:
fai_file = ref.fasta_idx(dd.get_ref_file(data))
cmd = ("{gsort} {in_file} {fai_file} | bedtools merge -i - -d {stats[merge_size]} | "
"bgzip -c > {tx_out_file}")
do.run(cmd.format(**locals()), "Prepare SV callable BED regions")
return vcfutils.bgzip_and_index(out_file, data["config"])
def split_vcf(in_file, ref_file, config, out_dir=None):
"""Split a VCF file into separate files by chromosome.
"""
if out_dir is None:
out_dir = os.path.join(os.path.dirname(in_file), "split")
out_files = []
with open(ref.fasta_idx(ref_file, config)) as in_handle:
for line in in_handle:
chrom, size = line.split()[:2]
out_file = os.path.join(out_dir,
os.path.basename(replace_suffix(append_stem(in_file, "-%s" % chrom), ".vcf")))
subset_vcf(in_file, (chrom, 0, size), out_file, config)
out_files.append(out_file)
return out_files
def subset_by_genes(in_file, data, out_dir, pad):
"""Subset BED file of regions to only those within pad of the final output.
"""
gene_file = regions.get_sv_bed(data, "exons", out_dir=os.path.dirname(in_file))
fai_file = ref.fasta_idx(dd.get_ref_file(data))
if not gene_file or not utils.file_exists(in_file):
return in_file
else:
out_file = os.path.join(out_dir, "%s-geneonly.bed" % utils.splitext_plus(os.path.basename(in_file))[0])
if not utils.file_uptodate(out_file, in_file):
with file_transaction(data, out_file) as tx_out_file:
want_region_file = "%s-targetregions%s" % utils.splitext_plus(out_file)
pybedtools.BedTool(gene_file).slop(g=fai_file, b=pad).merge().saveas(want_region_file)
pybedtools.BedTool(in_file).intersect(b=want_region_file).sort().saveas(tx_out_file)
return out_file
def add_genes(in_file, data, max_distance=10000, work_dir=None):
"""Add gene annotations to a BED file from pre-prepared RNA-seq data.
max_distance -- only keep annotations within this distance of event
"""
gene_file = regions.get_sv_bed(data, "exons", out_dir=os.path.dirname(in_file))
if gene_file and utils.file_exists(in_file):
out_file = "%s-annotated.bed" % utils.splitext_plus(in_file)[0]
if work_dir:
out_file = os.path.join(work_dir, os.path.basename(out_file))
if not utils.file_uptodate(out_file, in_file):
fai_file = ref.fasta_idx(dd.get_ref_file(data))
with file_transaction(data, out_file) as tx_out_file:
_add_genes_to_bed(in_file, gene_file, fai_file, tx_out_file, data, max_distance)
return out_file
else:
return in_file
def add_genes(in_file, data, max_distance=10000, work_dir=None):
"""Add gene annotations to a BED file from pre-prepared RNA-seq data.
max_distance -- only keep annotations within this distance of event
"""
gene_file = regions.get_sv_bed(data, "exons", out_dir=os.path.dirname(in_file))
if gene_file and utils.file_exists(in_file):
out_file = "%s-annotated.bed" % utils.splitext_plus(in_file)[0]
if work_dir:
out_file = os.path.join(work_dir, os.path.basename(out_file))
if not utils.file_uptodate(out_file, in_file):
fai_file = ref.fasta_idx(dd.get_ref_file(data))
with file_transaction(data, out_file) as tx_out_file:
_add_genes_to_bed(in_file, gene_file, fai_file, tx_out_file, data, max_distance)
return out_file
else:
return in_file
def split_vcf(in_file, ref_file, config, out_dir=None):
"""Split a VCF file into separate files by chromosome.
"""
if out_dir is None:
out_dir = os.path.join(os.path.dirname(in_file), "split")
out_files = []
with open(ref.fasta_idx(ref_file, config)) as in_handle:
for line in in_handle:
chrom, size = line.split()[:2]
out_file = os.path.join(
out_dir,
os.path.basename(
replace_suffix(append_stem(in_file, "-%s" % chrom),
".vcf")))
subset_vcf(in_file, (chrom, 0, size), out_file, config)
out_files.append(out_file)
return out_files
def add_genes(in_file, data, max_distance=10000):
"""Add gene annotations to a BED file from pre-prepared RNA-seq data.
max_distance -- only keep annotations within this distance of event
"""
gene_file = regions.get_sv_bed(data,
"exons",
out_dir=os.path.dirname(in_file))
if gene_file and utils.file_exists(in_file):
out_file = "%s-annotated.bed" % utils.splitext_plus(in_file)[0]
if not utils.file_uptodate(out_file, in_file):
input_rec = iter(pybedtools.BedTool(in_file)).next()
# keep everything after standard chrom/start/end, 1-based
extra_fields = range(4, len(input_rec.fields) + 1)
# keep the new gene annotation
gene_index = len(input_rec.fields) + 4
extra_fields.append(gene_index)
columns = ",".join([str(x) for x in extra_fields])
max_column = max(extra_fields) + 1
ops = ",".join(["distinct"] * len(extra_fields))
fai_file = ref.fasta_idx(dd.get_ref_file(data))
with file_transaction(data, out_file) as tx_out_file:
# swap over gene name to '.' if beyond maximum distance
# cut removes the last distance column which can cause issues
# with bedtools merge: 'ERROR: illegal character '.' found in integer conversion of string'
distance_filter = (
r"""awk -F$'\t' -v OFS='\t' '{if ($NF > %s) $%s = "."} {print}'"""
% (max_distance, gene_index))
sort_cmd = bedutils.get_sort_cmd()
cmd = (
"{sort_cmd} -k1,1 -k2,2n {in_file} | "
"bedtools closest -g <(cut -f1,2 {fai_file} | {sort_cmd} -k1,1 -k2,2n) "
"-d -t all -a - -b <({sort_cmd} -k1,1 -k2,2n {gene_file}) | "
"{distance_filter} | cut -f 1-{max_column} | "
"bedtools merge -i - -c {columns} -o {ops} -delim ',' > {tx_out_file}"
)
do.run(cmd.format(**locals()),
"Annotate BED file with gene info")
return out_file
else:
return in_file
def _get_coverage_file(in_bam, ref_file, region, region_file, depth, base_file, data):
"""Retrieve summary of coverage in a region.
Requires positive non-zero mapping quality at a position, matching GATK's
CallableLoci defaults.
"""
out_file = "%s-genomecov.bed" % utils.splitext_plus(base_file)[0]
if not utils.file_exists(out_file):
with file_transaction(data, out_file) as tx_out_file:
fai_file = ref.fasta_idx(ref_file, data["config"])
sambamba = config_utils.get_program("sambamba", data["config"])
bedtools = config_utils.get_program("bedtools", data["config"])
cmd = ("{sambamba} view -F 'mapping_quality > 0' -L {region_file} -f bam -l 1 {in_bam} | "
"{bedtools} genomecov -split -ibam stdin -bga -g {fai_file} "
"> {tx_out_file}")
do.run(cmd.format(**locals()), "bedtools genomecov: %s" % (str(region)), data)
# Empty output file, no coverage for the whole contig
if not utils.file_exists(out_file):
with file_transaction(data, out_file) as tx_out_file:
with open(tx_out_file, "w") as out_handle:
for feat in get_ref_bedtool(ref_file, data["config"], region):
out_handle.write("%s\t%s\t%s\t%s\n" % (feat.chrom, feat.start, feat.end, 0))
return out_file
