def run_rnaseq(bam_file, data, out_dir):
"""
Run qualimap for a rnaseq bam file and parse results
"""
strandedness = {"firststrand": "strand-specific-reverse",
"secondstrand": "strand-specific-forward",
"unstranded": "non-strand-specific"}
report_file = os.path.join(out_dir, "qualimapReport.html")
raw_file = os.path.join(out_dir, "rnaseq_qc_results.txt")
config = data["config"]
gtf_file = dd.get_gtf_file(data)
single_end = not bam.is_paired(bam_file)
library = strandedness[dd.get_strandedness(data)]
if not utils.file_exists(report_file):
utils.safe_makedir(out_dir)
bam.index(bam_file, config)
cmd = _rnaseq_qualimap_cmd(data, bam_file, out_dir, gtf_file, single_end, library)
do.run(cmd, "Qualimap for {}".format(dd.get_sample_name(data)))
cmd = "sed -i 's/bam file = .*/bam file = %s.bam/' %s" % (dd.get_sample_name(data), raw_file)
do.run(cmd, "Fix Name Qualimap for {}".format(dd.get_sample_name(data)))
metrics = _parse_rnaseq_qualimap_metrics(report_file)
metrics.update(_detect_duplicates(bam_file, out_dir, data))
metrics.update(_detect_rRNA(data))
metrics.update({"Average insert size": bam.estimate_fragment_size(bam_file)})
metrics = _parse_metrics(metrics)
return metrics
def _detect_rRNA(data, out_dir):
out_file = os.path.join(out_dir, "rRNA_metrics.txt")
if not utils.file_exists(out_file):
gtf_file = dd.get_gtf_file(data)
quant = tz.get_in(["quant", "tsv"], data)
if not quant:
salmon_dir = dd.get_salmon_dir(data)
if salmon_dir:
quant = os.path.join(salmon_dir, "quant", "quant.sf")
logger.info("Calculating RNA-seq rRNA metrics for %s." % quant)
rrna_features = gtf.get_rRNA(gtf_file)
transcripts = set([x[1] for x in rrna_features if x])
if not (transcripts and quant and utils.file_exists(quant)):
return {'rRNA': "NA", "rRNA_rate": "NA"}
sample_table = pd.read_csv(quant, sep="\t")
rrna_exp = list(map(float, sample_table[sample_table["Name"].isin(transcripts)]["NumReads"]))
total_exp = list(map(float, sample_table["NumReads"]))
rrna = sum(rrna_exp)
if sum(total_exp) == 0:
rrna_rate = "NA"
else:
rrna_rate = float(rrna) / sum(total_exp)
with file_transaction(out_file) as tx_out_file:
with open(tx_out_file, "w") as out_handle:
out_handle.write(",".join(["rRNA", str(rrna)]) + "\n")
out_handle.write(",".join(["rRNA_rate", str(rrna_rate)]) + "\n")
return _read_memoized_rrna(out_file)
def get_qc_tools(data):
"""Retrieve a list of QC tools to use based on configuration and analysis type.
Uses defaults if previously set.
"""
if dd.get_algorithm_qc(data):
return dd.get_algorithm_qc(data)
analysis = data["analysis"].lower()
to_run = []
if "fastqc" not in dd.get_tools_off(data):
to_run.append("fastqc")
if any([tool in dd.get_tools_on(data)
for tool in ["qualimap", "qualimap_full"]]):
to_run.append("qualimap")
if analysis.startswith("rna-seq"):
if gtf.is_qualimap_compatible(dd.get_gtf_file(data)):
to_run.append("qualimap_rnaseq")
else:
logger.debug("GTF not compatible with Qualimap, skipping.")
if analysis.startswith("smallrna-seq"):
to_run.append("small-rna")
if not analysis.startswith("smallrna-seq"):
to_run.append("samtools")
to_run.append("gemini")
if tz.get_in(["config", "algorithm", "kraken"], data):
to_run.append("kraken")
if analysis.startswith(("standard", "variant", "variant2")):
to_run += ["qsignature", "coverage", "variants", "picard"]
return to_run
def rnaseq_vardict_variant_calling(data):
sample = dd.get_sample_name(data)
variation_dir = os.path.join(dd.get_work_dir(data), "variation")
safe_makedir(variation_dir)
out_file = os.path.join(variation_dir, sample + "-vardict.vcf.gz")
if file_exists(out_file):
data = dd.set_vrn_file(data, out_file)
return data
vardict_cmd = vardict.get_vardict_command(data)
strandbias = "teststrandbias.R"
var2vcf = "var2vcf_valid.pl"
vcfstreamsort = config_utils.get_program("vcfstreamsort", data)
compress_cmd = "| bgzip -c"
freq = float(dd.get_min_allele_fraction(data, 20) / 100.0)
var2vcf_opts = "-v 50"
fix_ambig = vcfutils.fix_ambiguous_cl()
remove_dup = vcfutils.remove_dup_cl()
r_setup = ("unset R_HOME && export PATH=%s:$PATH && "
% os.path.dirname(Rscript_cmd()))
ref_file = dd.get_ref_file(data)
bamfile = dd.get_work_bam(data)
bed_file = gtf.gtf_to_bed(dd.get_gtf_file(data))
opts = " -c 1 -S 2 -E 3 -g 4 "
with file_transaction(out_file) as tx_out_file:
jvm_opts = vardict._get_jvm_opts(data, tx_out_file)
cmd = ("{r_setup}{jvm_opts}{vardict_cmd} -G {ref_file} -f {freq} "
"-N {sample} -b {bamfile} {opts} {bed_file} "
"| {strandbias}"
"| {var2vcf} -N {sample} -E -f {freq} {var2vcf_opts} "
"| {fix_ambig} | {remove_dup} | {vcfstreamsort} {compress_cmd} "
"> {tx_out_file}")
message = "Calling RNA-seq variants with VarDict"
do.run(cmd.format(**locals()), message)
data = dd.set_vrn_file(data, out_file)
return data
def _create_combined_fasta(data, out_dir):
"""
if there are genomes to be disambiguated, create a FASTA file of
all of the transcripts for all genomes
"""
items = disambiguate.split([data])
fasta_files = []
for i in items:
odata = i[0]
gtf_file = dd.get_gtf_file(odata)
ref_file = dd.get_ref_file(odata)
out_file = os.path.join(out_dir, dd.get_genome_build(odata) + ".fa")
if file_exists(out_file):
fasta_files.append(out_file)
else:
out_file = _gtf_to_fasta(gtf_file, ref_file, out_file)
out_file = _clean_gtf_fa(out_file, out_file)
fasta_files.append(out_file)
out_stem = os.path.join(out_dir, dd.get_genome_build(data))
if dd.get_disambiguate(data):
out_stem = "-".join([out_stem] + dd.get_disambiguate(data))
combined_file = out_stem + ".fa"
if file_exists(combined_file):
return combined_file
fasta_file_string = " ".join(fasta_files)
cmd = "cat {fasta_file_string} > {tx_out_file}"
with file_transaction(combined_file) as tx_out_file:
do.run(cmd.format(**locals()), "Combining transcriptome FASTA files.")
return combined_file
def _detect_rRNA(data):
gtf_file = dd.get_gtf_file(data)
count_file = dd.get_count_file(data)
rrna_features = gtf.get_rRNA(gtf_file)
genes = [x[0] for x in rrna_features if x]
count_table = pd.read_csv(count_file, sep="\t", names=["id", "counts"])
return {'rRNA': sum(count_table.ix[genes]["counts"])}
def count(data):
"""
count reads mapping to genes using featureCounts
http://subread.sourceforge.net
"""
in_bam = dd.get_work_bam(data)
gtf_file = dd.get_gtf_file(data)
work_dir = dd.get_work_dir(data)
out_dir = os.path.join(work_dir, "htseq-count")
safe_makedir(out_dir)
count_file = os.path.join(out_dir, dd.get_sample_name(data)) + ".counts"
if file_exists(count_file):
return count_file
featureCounts = config_utils.get_program("featureCounts", dd.get_config(data))
paired_flag = _paired_flag(in_bam)
strand_flag = _strand_flag(data)
cmd = ("{featureCounts} -a {gtf_file} -o {tx_count_file} -s {strand_flag} "
"{paired_flag} {in_bam}")
message = ("Count reads in {tx_count_file} mapping to {gtf_file} using "
"featureCounts")
with file_transaction(data, count_file) as tx_count_file:
do.run(cmd.format(**locals()), message.format(**locals()))
fixed_count_file = _format_count_file(count_file, data)
os.rename(fixed_count_file, count_file)
return count_file
def calling(data):
"""Main function to parallelize peak calling."""
chip_bam = dd.get_work_bam(data)
if data["work_bam_rep"] != "":
rep_bam = data["work_bam_rep"]
else:
rep_bam = ""
input_bam = data["work_bam_input"]
caller_fn = get_callers()[data["rmats_fn"]]
name = dd.get_sample_name(data)
fastq_file = fastq.get_fastq_files(data)
read_len = bam.fastq.estimate_read_length(fastq_file[0])
if read_len < 50:
read_len = 50
elif read_len > 50 and read_len < 75:
read_len = 75
else:
read_len = 100
if len(fastq_file) > 1:
read_pair = "paired"
else:
read_pair = "single"
out_dir = utils.safe_makedir(os.path.join(dd.get_work_dir(data), data["rmats_fn"], name ))
out_file = caller_fn(name, chip_bam, rep_bam, input_bam, dd.get_gtf_file(data), out_dir, read_len, read_pair, data["config"])
data["rmats_file"] = out_file
return [[data]]
def _get_ericscript_db(self, data):
transcript_file = dd.get_gtf_file(data)
if transcript_file and os.path.exists(transcript_file):
transcript_dir = os.path.dirname(transcript_file)
ericscript_dirs = glob.glob(os.path.join(transcript_dir, "ericscript", "ericscript_db*"))
if ericscript_dirs:
return sorted(ericscript_dirs)[-1]
def run_rnaseq(bam_file, data, out_dir):
"""
Run qualimap for a rnaseq bam file and parse results
"""
strandedness = {"firststrand": "strand-specific-reverse",
"secondstrand": "strand-specific-forward",
"unstranded": "non-strand-specific"}
# Qualimap results should be saved to a directory named after sample.
# MultiQC (for parsing additional data) picks the sample name after the dir as follows:
# <sample name>/raw_data_qualimapReport/insert_size_histogram.txt
results_dir = os.path.join(out_dir, dd.get_sample_name(data))
report_file = os.path.join(results_dir, "qualimapReport.html")
config = data["config"]
gtf_file = dd.get_gtf_file(data)
single_end = not bam.is_paired(bam_file)
library = strandedness[dd.get_strandedness(data)]
if not utils.file_exists(report_file):
with file_transaction(data, results_dir) as tx_out_dir:
utils.safe_makedir(tx_out_dir)
raw_file = os.path.join(tx_out_dir, "rnaseq_qc_results.txt")
bam.index(bam_file, config)
cmd = _rnaseq_qualimap_cmd(data, bam_file, tx_out_dir, gtf_file, single_end, library)
do.run(cmd, "Qualimap for {}".format(dd.get_sample_name(data)))
cmd = "sed -i 's/bam file = .*/bam file = %s.bam/' %s" % (dd.get_sample_name(data), raw_file)
do.run(cmd, "Fix Name Qualimap for {}".format(dd.get_sample_name(data)))
metrics = _parse_rnaseq_qualimap_metrics(report_file)
metrics.update(_detect_duplicates(bam_file, results_dir, data))
metrics.update(_detect_rRNA(data))
metrics.update({"Average_insert_size": bam.estimate_fragment_size(bam_file)})
metrics = _parse_metrics(metrics)
return metrics
def get_known_splicesites_file(align_dir, data):
gtf_file = dd.get_gtf_file(data)
splicesites = os.path.join(os.path.dirname(gtf_file),
"ref-transcripts-splicesites.txt")
if not file_exists(splicesites):
splicesites = create_splicesites_file(gtf_file, align_dir, data)
return splicesites
def count(data):
"""
count reads mapping to genes using featureCounts
http://subread.sourceforge.net
"""
in_bam = dd.get_work_bam(data)
sorted_bam = bam.sort(in_bam, dd.get_config(data), order="queryname")
gtf_file = dd.get_gtf_file(data)
work_dir = dd.get_work_dir(data)
out_dir = os.path.join(work_dir, "htseq-count")
safe_makedir(out_dir)
count_file = os.path.join(out_dir, dd.get_sample_name(data)) + ".counts"
summary_file = os.path.join(out_dir, dd.get_sample_name(data)) + ".counts.summary"
if file_exists(count_file):
return count_file
featureCounts = config_utils.get_program("featureCounts", dd.get_config(data))
paired_flag = _paired_flag(in_bam)
strand_flag = _strand_flag(data)
filtered_bam = bam.filter_primary(sorted_bam, data)
cmd = "{featureCounts} -a {gtf_file} -o {tx_count_file} -s {strand_flag} " "{paired_flag} {filtered_bam}"
message = "Count reads in {tx_count_file} mapping to {gtf_file} using " "featureCounts"
with file_transaction(data, [count_file, summary_file]) as tx_files:
tx_count_file, tx_summary_file = tx_files
do.run(cmd.format(**locals()), message.format(**locals()))
fixed_count_file = _format_count_file(count_file, data)
fixed_summary_file = _change_sample_name(summary_file, dd.get_sample_name(data), data=data)
shutil.move(fixed_count_file, count_file)
shutil.move(fixed_summary_file, summary_file)
return count_file
def run_stringtie_expression(data):
"""
estimate expression from Stringtie, using the bcbio datadict
does not do transcriptome assembly
"""
bam = dd.get_work_bam(data)
gtf_file = dd.get_gtf_file(data)
num_cores = dd.get_num_cores(data)
sample_name = dd.get_sample_name(data)
out_dir = os.path.join("stringtie", sample_name)
isoform_fpkm = os.path.join(out_dir, sample_name + ".isoform.fpkm")
gene_fpkm = os.path.join(out_dir, sample_name + ".fpkm")
if file_exists(isoform_fpkm) and file_exists(gene_fpkm):
data = dd.set_cufflinks_dir(data, out_dir)
data = dd.set_fpkm(data, gene_fpkm)
data = dd.set_fpkm_isoform(data, isoform_fpkm)
return data
with file_transaction(data, out_dir) as tx_out_dir:
exon_file, transcript_file = _stringtie_expression(bam, gtf_file, num_cores, tx_out_dir)
df = _parse_ballgown(transcript_file)
_write_fpkms(df, tx_out_dir, sample_name)
data = dd.set_cufflinks_dir(data, out_dir)
data = dd.set_fpkm(data, gene_fpkm)
data = dd.set_fpkm_isoform(data, isoform_fpkm)
return data
def run(data):
"""Quantitaive isoforms expression by eXpress"""
name = dd.get_sample_name(data)
in_bam = dd.get_transcriptome_bam(data)
config = data['config']
if not in_bam:
logger.info("Transcriptome-mapped BAM file not found, skipping eXpress.")
return data
gtf_fasta = gtf.gtf_to_fasta(dd.get_gtf_file(data), dd.get_ref_file(data))
out_dir = os.path.join(dd.get_work_dir(data), "express", name)
out_file = os.path.join(out_dir, name + ".xprs")
express = config_utils.get_program("express", data['config'])
strand = _set_stranded_flag(in_bam, data)
if not file_exists(out_file):
with tx_tmpdir(data) as tmp_dir:
with file_transaction(out_dir) as tx_out_dir:
bam_file = _prepare_bam_file(in_bam, tmp_dir, config)
cmd = ("{express} --no-update-check -o {tx_out_dir} {strand} {gtf_fasta} {bam_file}")
do.run(cmd.format(**locals()), "Run express on %s." % in_bam, {})
shutil.move(os.path.join(out_dir, "results.xprs"), out_file)
eff_count_file = _get_column(out_file, out_file.replace(".xprs", "_eff.counts"), 7)
tpm_file = _get_column(out_file, out_file.replace("xprs", "tpm"), 14)
fpkm_file = _get_column(out_file, out_file.replace("xprs", "fpkm"), 10)
data = dd.set_express_counts(data, eff_count_file)
data = dd.set_express_tpm(data, tpm_file)
data = dd.set_express_fpkm(data, fpkm_file)
return data
def align_transcriptome(fastq_file, pair_file, ref_file, data):
"""
bwa mem with settings for aligning to the transcriptome for eXpress/RSEM/etc
"""
work_bam = dd.get_work_bam(data)
base, ext = os.path.splitext(work_bam)
out_file = base + ".transcriptome" + ext
if utils.file_exists(out_file):
data = dd.set_transcriptome_bam(data, out_file)
return data
# bwa mem needs phred+33 quality, so convert if it is Illumina
if dd.get_quality_format(data).lower() == "illumina":
logger.info("bwa mem does not support the phred+64 quality format, " "converting %s and %s to phred+33.")
fastq_file = fastq.groom(fastq_file, in_qual="fastq-illumina", data=data)
if pair_file:
pair_file = fastq.groom(pair_file, in_qual="fastq-illumina", data=data)
bwa = config_utils.get_program("bwa", data["config"])
gtf_file = dd.get_gtf_file(data)
gtf_fasta = index_transcriptome(gtf_file, ref_file, data)
args = " ".join(_bwa_args_from_config(data["config"]))
num_cores = data["config"]["algorithm"].get("num_cores", 1)
cmd = (
"{bwa} mem {args} -a -t {num_cores} {gtf_fasta} {fastq_file} "
"{pair_file} | samtools view -bhS - > {tx_out_file}"
)
with file_transaction(out_file) as tx_out_file:
message = "Aligning %s and %s to the transcriptome." % (fastq_file, pair_file)
do.run(cmd.format(**locals()), message)
data = dd.set_transcriptome_bam(data, out_file)
return data
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, dd.get_lane(data)) + ".bam"
if file_exists(out_file):
data = dd.set_work_bam(data, out_file)
return data
cmd = ("{hisat2} -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 can use
if dd.get_assemble_transcripts(data):
cmd += "--dta-cufflinks "
if dd.get_analysis(data).lower() == "rna-seq":
gtf_file = dd.get_gtf_file(data)
splicesites = os.path.join(os.path.dirname(gtf_file),
"ref-transcripts-splicesites.txt")
cmd += "--known-splicesite-infile {splicesites} "
message = "Aligning %s and %s with hisat2." %(fastq_file, pair_file)
with file_transaction(out_file) as tx_out_file:
cmd += " | " + postalign.sam_to_sortbam_cl(data, tx_out_file)
do.run(cmd.format(**locals()), message)
data = dd.set_work_bam(data, out_file)
return data
def combine_files(samples):
"""
after quantitation, combine the counts/FPKM/TPM/etc into a single table with
all samples
"""
gtf_file = dd.get_gtf_file(samples[0][0], None)
dexseq_gff = dd.get_dexseq_gff(samples[0][0])
# combine featureCount files
count_files = filter_missing([dd.get_count_file(x[0]) for x in samples])
combined = count.combine_count_files(count_files, ext=".counts")
annotated = count.annotate_combined_count_file(combined, gtf_file)
# combine eXpress files
express_counts_combined = combine_express(samples, combined)
# combine Cufflinks files
fpkm_combined_file = os.path.splitext(combined)[0] + ".fpkm"
fpkm_files = filter_missing([dd.get_fpkm(x[0]) for x in samples])
if fpkm_files:
fpkm_combined = count.combine_count_files(fpkm_files, fpkm_combined_file)
else:
fpkm_combined = None
fpkm_isoform_combined_file = os.path.splitext(combined)[0] + ".isoform.fpkm"
isoform_files = filter_missing([dd.get_fpkm_isoform(x[0]) for x in samples])
if isoform_files:
fpkm_isoform_combined = count.combine_count_files(isoform_files,
fpkm_isoform_combined_file,
".isoform.fpkm")
else:
fpkm_isoform_combined = None
# combine DEXseq files
dexseq_combined_file = os.path.splitext(combined)[0] + ".dexseq"
to_combine_dexseq = filter_missing([dd.get_dexseq_counts(data[0]) for data in samples])
if to_combine_dexseq:
dexseq_combined = count.combine_count_files(to_combine_dexseq,
dexseq_combined_file, ".dexseq")
dexseq.create_dexseq_annotation(dexseq_gff, dexseq_combined)
else:
dexseq_combined = None
samples = spikein.combine_spikein(samples)
updated_samples = []
for data in dd.sample_data_iterator(samples):
data = dd.set_combined_counts(data, combined)
if annotated:
data = dd.set_annotated_combined_counts(data, annotated)
if fpkm_combined:
data = dd.set_combined_fpkm(data, fpkm_combined)
if fpkm_isoform_combined:
data = dd.set_combined_fpkm_isoform(data, fpkm_isoform_combined)
if express_counts_combined:
data = dd.set_express_counts(data, express_counts_combined['counts'])
data = dd.set_express_tpm(data, express_counts_combined['tpm'])
data = dd.set_express_fpkm(data, express_counts_combined['fpkm'])
data = dd.set_isoform_to_gene(data, express_counts_combined['isoform_to_gene'])
if dexseq_combined:
data = dd.set_dexseq_counts(data, dexseq_combined_file)
updated_samples.append([data])
return updated_samples
def align(fastq_file, pair_file, ref_file, names, align_dir, data):
if not ref_file:
logger.error("STAR index not found. We don't provide the STAR indexes "
"by default because they are very large. You can install "
"the index for your genome with: bcbio_nextgen.py upgrade "
"--aligners star --genomes genome-build-name --data")
sys.exit(1)
max_hits = 10
srna = True if data["analysis"].lower().startswith("smallrna-seq") else False
srna_opts = ""
if srna:
max_hits = 1000
srna_opts = "--alignIntronMax 1"
config = data["config"]
star_dirs = _get_star_dirnames(align_dir, data, names)
if file_exists(star_dirs.final_out):
data = _update_data(star_dirs.final_out, star_dirs.out_dir, names, data)
return data
star_path = config_utils.get_program("STAR", config)
fastq_files = " ".join([fastq_file, pair_file]) if pair_file else fastq_file
num_cores = dd.get_num_cores(data)
gtf_file = dd.get_gtf_file(data)
if ref_file.endswith("chrLength"):
ref_file = os.path.dirname(ref_file)
with file_transaction(data, align_dir) as tx_align_dir:
tx_star_dirnames = _get_star_dirnames(tx_align_dir, data, names)
tx_out_dir, tx_out_file, tx_out_prefix, tx_final_out = tx_star_dirnames
safe_makedir(tx_align_dir)
safe_makedir(tx_out_dir)
cmd = ("{star_path} --genomeDir {ref_file} --readFilesIn {fastq_files} "
"--runThreadN {num_cores} --outFileNamePrefix {tx_out_prefix} "
"--outReadsUnmapped Fastx --outFilterMultimapNmax {max_hits} "
"--outStd SAM {srna_opts} "
"--outSAMunmapped Within --outSAMattributes %s " % " ".join(ALIGN_TAGS))
cmd += _add_sj_index_commands(fastq_file, ref_file, gtf_file) if not srna else ""
cmd += " --readFilesCommand zcat " if is_gzipped(fastq_file) else ""
cmd += _read_group_option(names)
fusion_mode = utils.get_in(data, ("config", "algorithm", "fusion_mode"), False)
if fusion_mode:
cmd += (" --chimSegmentMin 12 --chimJunctionOverhangMin 12 "
"--chimScoreDropMax 30 --chimSegmentReadGapMax 5 "
"--chimScoreSeparation 5 "
"--chimOutType WithinSAM ")
strandedness = utils.get_in(data, ("config", "algorithm", "strandedness"),
"unstranded").lower()
if strandedness == "unstranded" and not srna:
cmd += " --outSAMstrandField intronMotif "
if not srna:
cmd += " --quantMode TranscriptomeSAM "
cmd += " | " + postalign.sam_to_sortbam_cl(data, tx_final_out)
run_message = "Running STAR aligner on %s and %s" % (fastq_file, ref_file)
do.run(cmd.format(**locals()), run_message, None)
print("hello")
data = _update_data(star_dirs.final_out, star_dirs.out_dir, names, data)
return data
def run_salmon_index(*samples):
for data in dd.sample_data_iterator(samples):
work_dir = dd.get_work_dir(data)
salmon_dir = os.path.join(work_dir, "salmon")
gtf_file = dd.get_gtf_file(data)
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
salmon_index(gtf_file, fasta_file, data, salmon_dir)
return samples
def run_rapmap_index(*samples):
for data in dd.sample_data_iterator(samples):
work_dir = dd.get_work_dir(data)
rapmap_dir = os.path.join(work_dir, "rapmap")
gtf_file = dd.get_gtf_file(data)
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
rapmap_index(gtf_file, fasta_file, "quasi", data, rapmap_dir)
return samples
def tagcount(data):
bam = dd.get_transcriptome_bam(data)
umi_dir = os.path.join(dd.get_work_dir(data), "umis")
sample_dir = os.path.join(umi_dir, dd.get_sample_name(data))
out_prefix = os.path.join(sample_dir, dd.get_sample_name(data))
out_file = out_prefix + ".mtx"
if file_exists(out_file):
data = dd.set_count_file(data, out_file)
return [[data]]
umis = config_utils.get_program("umis", data, default="umis")
safe_makedir(sample_dir)
cutoff = dd.get_minimum_barcode_depth(data)
cb_histogram = os.path.join(sample_dir, "cb-histogram.txt")
positional = "--positional" if dd.get_positional_umi(data, False) else ""
if use_installed_transcriptome(data):
gtf_file = dd.get_gtf_file(data)
else:
gtf_file = dd.get_transcriptome_gtf(data, None)
if gtf_file:
gene_map_file = os.path.join(dd.get_work_dir(data), "annotation",
os.path.splitext(gtf_file)[0] + "-tx2gene.tsv")
gene_map_file = gtf.tx2genefile(gtf_file, gene_map_file, tsv=True)
gene_map_flag = " --genemap {0} ".format(gene_map_file)
else:
gene_map_flag = ""
message = "Counting alignments of transcripts in %s." % bam
cmd = ("{umis} fasttagcount --cb_cutoff {cutoff} "
"{gene_map_flag} "
"{positional} "
"--cb_histogram {cb_histogram}")
out_files = [out_file, out_file + ".rownames", out_file + ".colnames"]
umi_matrix_file = out_prefix + "-dupes.mtx"
out_files += [umi_matrix_file, umi_matrix_file + ".rownames",
umi_matrix_file + ".colnames"]
if has_umi_matrix(data):
umi_matrix_flag = " --umi_matrix {tx_umi_matrix_full} "
else:
umi_matrix_flag = ""
cmd += umi_matrix_flag
cmd += " {bam} {tx_out_file_full}"
with file_transaction(out_files) as tx_out_files:
tx_out_file = tx_out_files[0]
tx_out_file_full = tx_out_file + ".full"
tx_umi_matrix = tx_out_files[3]
tx_umi_matrix_full = tx_out_files[3] + ".full"
do.run(cmd.format(**locals()), message)
cmd = ("{umis} sparse {tx_out_file_full} {tx_out_file}")
message = "Converting %s to sparse format." % tx_out_file_full
do.run(cmd.format(**locals()), message)
if has_umi_matrix(data):
cmd = ("{umis} sparse {tx_umi_matrix_full} {tx_umi_matrix}")
message = "Converting %s to sparse format." % tx_umi_matrix_full
do.run(cmd.format(**locals()), message)
data = dd.set_count_file(data, out_file)
return [[data]]
def run_kallisto_index(*samples):
for data in dd.sample_data_iterator(samples):
work_dir = dd.get_work_dir(data)
kallisto_dir = os.path.join(work_dir, "kallisto")
gtf_file = dd.get_gtf_file(data)
assert file_exists(gtf_file), "%s was not found, exiting." % gtf_file
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
kallisto_index(gtf_file, fasta_file, data, kallisto_dir)
return samples
def _get_files(data):
mapped = bam.mapped(data["work_bam"], data["config"])
in_file = bam.sort(mapped, data["config"], order="queryname")
gtf_file = dd.get_gtf_file(data)
work_dir = dd.get_work_dir(data)
out_dir = os.path.join(work_dir, "htseq-count")
sample_name = dd.get_sample_name(data)
out_file = os.path.join(out_dir, sample_name + ".counts")
stats_file = os.path.join(out_dir, sample_name + ".stats")
return in_file, gtf_file, out_file, stats_file
def align(fastq_file, pair_file, ref_file, names, align_dir, data):
max_hits = 10
srna = True if data["analysis"].lower().startswith("smallrna-seq") else False
srna_opts = ""
if srna:
max_hits = 1000
srna_opts = "--alignIntronMax 1"
config = data["config"]
out_prefix = os.path.join(align_dir, dd.get_lane(data))
out_file = out_prefix + "Aligned.out.sam"
out_dir = os.path.join(align_dir, "%s_star" % dd.get_lane(data))
if not ref_file:
logger.error("STAR index not found. We don't provide the STAR indexes "
"by default because they are very large. You can install "
"the index for your genome with: bcbio_nextgen.py upgrade "
"--aligners star --genomes genome-build-name --data")
sys.exit(1)
final_out = os.path.join(out_dir, "{0}.bam".format(names["sample"]))
if file_exists(final_out):
data = _update_data(final_out, out_dir, names, data)
return data
star_path = config_utils.get_program("STAR", config)
fastq_files = " ".join([fastq_file, pair_file]) if pair_file else fastq_file
num_cores = dd.get_num_cores(data)
gtf_file = dd.get_gtf_file(data)
safe_makedir(align_dir)
cmd = ("{star_path} --genomeDir {ref_file} --readFilesIn {fastq_files} "
"--runThreadN {num_cores} --outFileNamePrefix {out_prefix} "
"--outReadsUnmapped Fastx --outFilterMultimapNmax {max_hits} "
"--outStd SAM {srna_opts} "
"--outSAMunmapped Within --outSAMattributes %s " % " ".join(ALIGN_TAGS))
cmd += _add_sj_index_commands(fastq_file, ref_file, gtf_file)
cmd += " --readFilesCommand zcat " if is_gzipped(fastq_file) else ""
cmd += _read_group_option(names)
fusion_mode = utils.get_in(data, ("config", "algorithm", "fusion_mode"), False)
if fusion_mode:
cmd += " --chimSegmentMin 15 --chimJunctionOverhangMin 15"
strandedness = utils.get_in(data, ("config", "algorithm", "strandedness"),
"unstranded").lower()
if strandedness == "unstranded" and not srna:
cmd += " --outSAMstrandField intronMotif "
if dd.get_transcriptome_align(data) and not is_transcriptome_broken(data):
cmd += " --quantMode TranscriptomeSAM "
with file_transaction(data, final_out) as tx_final_out:
cmd += " | " + postalign.sam_to_sortbam_cl(data, tx_final_out)
run_message = "Running STAR aligner on %s and %s" % (fastq_file, ref_file)
do.run(cmd.format(**locals()), run_message, None)
data = _update_data(final_out, out_dir, names, data)
return data
def _detect_rRNA(data):
gtf_file = dd.get_gtf_file(data)
count_file = dd.get_count_file(data)
rrna_features = gtf.get_rRNA(gtf_file)
genes = [x[0] for x in rrna_features if x]
if not genes:
return {'rRNA': "NA", "rRNA_rate": "NA"}
count_table = pd.read_csv(count_file, sep="\t", names=["id", "counts"])
rrna = sum(count_table[count_table["id"].isin(genes)]["counts"])
rrna_rate = float(rrna) / sum(count_table["counts"])
return {'rRNA': str(rrna), 'rRNA_rate': str(rrna_rate)}
def run_sailfish_index(*samples):
samples = [utils.to_single_data(x) for x in samples]
Build = namedtuple('Build', ['build', 'ref', 'gtf'])
builds = {Build(get_build_string(x), dd.get_ref_file(x), dd.get_gtf_file(x))
for x in samples}
data = samples[0]
indexdirs = {}
for build in builds:
indexdirs[build.build] = sailfish_index(build.ref, build.gtf, data,
build.build)
return [[x] for x in samples]
def run_salmon_bam(data):
samplename = dd.get_sample_name(data)
work_dir = dd.get_work_dir(data)
salmon_dir = os.path.join(work_dir, "salmon", samplename)
gtf_file = dd.get_gtf_file(data)
bam_file = dd.get_transcriptome_bam(data)
fasta_file = dd.get_ref_file(data)
out_file = salmon_quant_bam(bam_file, salmon_dir, gtf_file, fasta_file, data)
data = dd.set_salmon(data, out_file)
data = dd.set_salmon_dir(data, salmon_dir)
data = dd.set_salmon_fraglen_file(data, _get_fraglen_file(salmon_dir))
return [[data]]
def run_sailfish_index(*samples):
fq1, _ = dd.get_input_sequence_files(samples[0][0])
kmer_size = estimate_kmer_size(fq1)
Build = namedtuple('Build', ['build', 'ref', 'gtf'])
builds = {Build(get_build_string(x), dd.get_ref_file(x), dd.get_gtf_file(x))
for x in dd.sample_data_iterator(samples)}
data = samples[0][0]
indexdirs = {}
for build in builds:
indexdirs[build.build] = sailfish_index(build.ref, build.gtf, data,
build.build, kmer_size)
return samples
def stringtie_merge(*samples):
to_merge = filter_missing(flatten([dd.get_assembled_gtf(data) for data in
dd.sample_data_iterator(samples)]))
data = samples[0][0]
ref_file = dd.get_sam_ref(data)
gtf_file = dd.get_gtf_file(data)
num_cores = dd.get_num_cores(data)
merged_gtf = stringtie.merge(to_merge, ref_file, gtf_file, num_cores, data)
updated_samples = []
for data in dd.sample_data_iterator(samples):
data = dd.set_merged_gtf(data, merged_gtf)
updated_samples.append([data])
return updated_samples
def run_salmon_bam(data):
samplename = dd.get_sample_name(data)
work_dir = dd.get_work_dir(data)
salmon_dir = os.path.join(work_dir, "salmon", samplename)
gtf_file = dd.get_gtf_file(data)
bam_file = dd.get_transcriptome_bam(data)
assert file_exists(gtf_file), "%s was not found, exiting." % gtf_file
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
out_file = salmon_quant_bam(bam_file, salmon_dir, gtf_file, fasta_file, data)
data = dd.set_salmon(data, out_file)
data = dd.set_salmon_dir(data, salmon_dir)
return [[data]]
def index(ref_file, out_dir, data):
"""Create a bismark index in the defined reference directory.
"""
(ref_dir, local_file) = os.path.split(ref_file)
gtf_file = dd.get_gtf_file(data)
bismark = config_utils.find_program("bismark", data["config"])
if not utils.file_exists(gtf_file):
raise ValueError("%s not found, could not create a star index." %
(gtf_file))
if not utils.file_exists(out_dir):
with tx_tmpdir(data, os.path.dirname(out_dir)) as tx_out_dir:
num_cores = dd.get_cores(data)
cmd = "{bismark} --bowtie2 -p {num_cores} -n 1 -o {tx_out_dir} --basename {sample} --unmapped {ref_file} {in_fastq}"
do.run(cmd.format(**locals()), "Index STAR")
if os.path.exists(out_dir):
shutil.rmtree(out_dir)
shutil.move(tx_out_dir, out_dir)
return out_dir
def run_rapmap_pseudoalign(data):
samplename = dd.get_sample_name(data)
files = dd.get_input_sequence_files(data)
work_dir = dd.get_work_dir(data)
if len(files) == 2:
fq1, fq2 = files
else:
fq1, fq2 = files[0], None
rapmap_dir = os.path.join(work_dir, "rapmap", samplename)
gtf_file = dd.get_gtf_file(data)
assert file_exists(gtf_file), "%s was not found, exiting." % gtf_file
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
out_file = rapmap_pseudoalign(fq1, fq2, rapmap_dir, gtf_file, fasta_file,
data)
data = dd.set_work_bam(data, out_file)
data = dd.set_transcriptome_bam(data, out_file)
return data
def run_sailfish(data):
samplename = dd.get_sample_name(data)
files = dd.get_input_sequence_files(data)
work_dir = dd.get_work_dir(data)
if len(files) == 2:
fq1, fq2 = files
else:
fq1, fq2 = files[0], None
sailfish_dir = os.path.join(work_dir, "sailfish", samplename)
gtf_file = dd.get_gtf_file(data)
assert file_exists(gtf_file), "%s was not found, exiting." % gtf_file
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
stranded = dd.get_strandedness(data).lower()
out_file = sailfish(fq1, fq2, sailfish_dir, gtf_file, fasta_file, stranded, data)
data = dd.set_sailfish(data, out_file)
data = dd.set_sailfish_dir(data, sailfish_dir)
return [[data]]
def _detect_rRNA(data):
sample = dd.get_sample_name(data)
gtf_file = dd.get_gtf_file(data)
tidy_file = dd.get_sailfish_tidy(data)
rrna_features = gtf.get_rRNA(gtf_file)
transcripts = set([x[1] for x in rrna_features if x])
if not (transcripts and tidy_file):
return {'rRNA': "NA", "rRNA_rate": "NA"}
count_table = pd.read_csv(tidy_file, sep="\t")
sample_table = count_table[count_table["sample"].isin([sample])]
rrna_exp = map(
float, sample_table[sample_table["id"].isin(transcripts)]["numreads"])
total_exp = map(float, sample_table["numreads"])
rrna = sum(rrna_exp)
if sum(total_exp) == 0:
return {'rRNA': str(rrna), 'rRNA_rate': "NA"}
rrna_rate = float(rrna) / sum(total_exp)
return {'rRNA': str(rrna), 'rRNA_rate': str(rrna_rate)}
def cufflinks_merge(*samples):
to_merge = set(
filter_missing(
flatten([
dd.get_assembled_gtf(data)
for data in dd.sample_data_iterator(samples)
])))
data = samples[0][0]
ref_file = dd.get_sam_ref(data)
gtf_file = dd.get_gtf_file(data)
num_cores = dd.get_num_cores(data)
merged_gtf = cufflinks.merge(to_merge, ref_file, gtf_file, num_cores,
samples[0][0])
updated_samples = []
for data in dd.sample_data_iterator(samples):
data = dd.set_merged_gtf(data, merged_gtf)
updated_samples.append([data])
return updated_samples
def index(ref_file, out_dir, data):
"""Create a STAR index in the defined reference directory.
"""
(ref_dir, local_file) = os.path.split(ref_file)
gtf_file = dd.get_gtf_file(data)
if not utils.file_exists(gtf_file):
raise ValueError("%s not found, could not create a star index." % (gtf_file))
if not utils.file_exists(out_dir):
with tx_tmpdir(data, os.path.dirname(out_dir)) as tx_out_dir:
num_cores = dd.get_cores(data)
cmd = ("STAR --genomeDir {tx_out_dir} --genomeFastaFiles {ref_file} "
"--runThreadN {num_cores} "
"--runMode genomeGenerate --sjdbOverhang 99 --sjdbGTFfile {gtf_file}")
do.run(cmd.format(**locals()), "Index STAR")
if os.path.exists(out_dir):
shutil.rmtree(out_dir)
shutil.move(tx_out_dir, out_dir)
return out_dir
def run_salmon_reads(data):
data = utils.to_single_data(data)
samplename = dd.get_sample_name(data)
work_dir = dd.get_work_dir(data)
salmon_dir = os.path.join(work_dir, "salmon", samplename)
gtf_file = dd.get_gtf_file(data)
files = dd.get_input_sequence_files(data)
if len(files) == 2:
fq1, fq2 = files
else:
fq1, fq2 = files[0], None
assert file_exists(gtf_file), "%s was not found, exiting." % gtf_file
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
out_file = salmon_quant_reads(fq1, fq2, salmon_dir, gtf_file, fasta_file, data)
data = dd.set_sailfish(data, out_file)
data = dd.set_sailfish_dir(data, salmon_dir)
return [[data]]
def run_ataqv(data):
if not dd.get_chip_method(data) == "atac":
return None
work_dir = dd.get_work_dir(data)
sample_name = dd.get_sample_name(data)
out_dir = os.path.join(work_dir, "qc", sample_name, "ataqv")
peak_file = get_full_peaks(data)
bam_file = get_unfiltered_bam(data)
out_file = os.path.join(out_dir, sample_name + ".ataqv.json.gz")
if not peak_file:
logger.info(f"Full peak file for {sample_name} not found, skipping ataqv")
return None
if not bam_file:
logger.info(f"Unfiltered BAM file for {sample_name} not found, skipping ataqv")
return None
if utils.file_exists(out_file):
return out_file
tss_bed_file = os.path.join(out_dir, "TSS.bed")
tss_bed_file = gtf.get_tss_bed(dd.get_gtf_file(data), tss_bed_file, data, padding=0)
if chromhacks.is_human(data):
organism = "human"
autosomal_reference_flag = ""
elif chromhacks.is_mouse(data):
organism = "mouse"
autosomal_reference_flag = ""
else:
autosomal_reference = os.path.join(out_dir, "autosomal.txt")
autosomal_reference = _make_autosomal_reference_file(autosomal_reference, data)
organism = "None"
autosomal_reference_flag = f"--autosomal-reference-file {autosomal_reference} "
ataqv = config_utils.get_program("ataqv", data)
mitoname = chromhacks.get_mitochondrial_chroms(data)[0]
if not ataqv:
logger.info(f"ataqv executable not found, skipping running ataqv.")
return None
with file_transaction(out_file) as tx_out_file:
cmd = (f"{ataqv} --peak-file {peak_file} --name {sample_name} --metrics-file {tx_out_file} "
f"--tss-file {tss_bed_file} {autosomal_reference_flag} "
f"--ignore-read-groups --mitochondrial-reference-name {mitoname} "
f"--tss-extension 1000 "
f"{organism} {bam_file}")
message = f"Running ataqv on {sample_name}."
do.run(cmd, message)
return out_file
def get_qc_tools(data):
"""Retrieve a list of QC tools to use based on configuration and analysis type.
Uses defaults if previously set.
"""
if dd.get_algorithm_qc(data):
return dd.get_algorithm_qc(data)
analysis = data["analysis"].lower()
to_run = []
if tz.get_in(["config", "algorithm", "kraken"], data):
to_run.append("kraken")
if "fastqc" not in dd.get_tools_off(data):
to_run.append("fastqc")
if any([
tool in dd.get_tools_on(data)
for tool in ["qualimap", "qualimap_full"]
]):
to_run.append("qualimap")
if analysis.startswith("rna-seq") or analysis == "smallrna-seq":
if "qualimap" not in dd.get_tools_off(data):
if gtf.is_qualimap_compatible(dd.get_gtf_file(data)):
to_run.append("qualimap_rnaseq")
else:
logger.debug("GTF not compatible with Qualimap, skipping.")
if analysis.startswith("smallrna-seq"):
to_run.append("small-rna")
to_run.append("atropos")
if "coverage_qc" not in dd.get_tools_off(data):
to_run.append("samtools")
if analysis.startswith(("standard", "variant", "variant2")):
if "coverage_qc" not in dd.get_tools_off(data):
to_run += ["coverage", "picard"]
to_run += ["qsignature", "variants"]
if peddy.is_human(data):
to_run += ["peddy"]
if vcfutils.get_paired([data]):
to_run += ["viral"]
if damage.should_filter([data]):
to_run += ["damage"]
if dd.get_umi_consensus(data):
to_run += ["umi"]
if tz.get_in(["config", "algorithm", "preseq"], data):
to_run.append("preseq")
return to_run
def run_kallisto_rnaseq(data):
samplename = dd.get_sample_name(data)
work_dir = dd.get_work_dir(data)
kallisto_dir = os.path.join(work_dir, "kallisto", samplename)
gtf_file = dd.get_gtf_file(data)
files = dd.get_input_sequence_files(data)
if len(files) == 2:
fq1, fq2 = files
else:
fq1, fq2 = files[0], None
assert file_exists(gtf_file), "%s was not found, exiting." % gtf_file
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
assert fq2, ("bcbio doesn't support kallisto for single-end reads, we can "
"add support for this if you open up an issue about it here: "
"https://github.com/bcbio/bcbio-nextgen/issues")
out_file = kallisto_rnaseq(fq1, fq2, kallisto_dir, gtf_file, fasta_file, data)
data = dd.set_kallisto_quant(data, out_file)
return [[data]]
def run_kallisto_rnaseq(data):
samplename = dd.get_sample_name(data)
work_dir = dd.get_work_dir(data)
kallisto_dir = os.path.join(work_dir, "kallisto", samplename)
gtf_file = dd.get_transcriptome_gtf(data, default=dd.get_gtf_file(data))
files = dd.get_input_sequence_files(data)
if len(files) == 2:
fq1, fq2 = files
else:
fq1, fq2 = files[0], None
assert file_exists(gtf_file), "%s was not found, exiting." % gtf_file
fasta_file = dd.get_ref_file(data)
assert file_exists(fasta_file), "%s was not found, exiting." % fasta_file
assert fq2, ("We don't support kallisto for single-end reads and fusion "
"calling with pizzly does not accept single end reads.")
out_file = kallisto_rnaseq(fq1, fq2, kallisto_dir, gtf_file, fasta_file,
data)
data = dd.set_kallisto_quant(data, out_file)
return [[data]]
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, dd.get_lane(data)) + ".bam"
if file_exists(out_file):
data = dd.set_work_bam(data, out_file)
return data
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":
gtf_file = dd.get_gtf_file(data)
splicesites = os.path.join(os.path.dirname(gtf_file),
"ref-transcripts-splicesites.txt")
if not file_exists(splicesites):
splicesites = create_splicesites_file(gtf_file, align_dir, data)
# empty splicesite files means there is no splicing, so skip this option
# if there is no splicing for this organism
if file_exists(splicesites):
cmd += "--known-splicesite-infile {splicesites} "
# apply additional hisat2 options
cmd += " ".join(_get_options_from_config(data))
message = "Aligning %s and %s with hisat2." % (fastq_file, pair_file)
with file_transaction(data, out_file) as tx_out_file:
cmd += " | " + postalign.sam_to_sortbam_cl(data, tx_out_file)
do.run(cmd.format(**locals()), message)
data = dd.set_work_bam(data, out_file)
return data
def _detect_rRNA(data):
sample = dd.get_sample_name(data)
gtf_file = dd.get_gtf_file(data)
salmon_dir = dd.get_salmon_dir(data)
quant = os.path.join(salmon_dir, "quant", "quant.sf")
rrna_features = gtf.get_rRNA(gtf_file)
transcripts = set([x[1] for x in rrna_features if x])
if not (transcripts and utils.file_exists(quant)):
return {'rRNA': "NA", "rRNA_rate": "NA"}
sample_table = pd.read_csv(quant, sep="\t")
rrna_exp = map(
float,
sample_table[sample_table["Name"].isin(transcripts)]["NumReads"])
total_exp = map(float, sample_table["NumReads"])
rrna = sum(rrna_exp)
if sum(total_exp) == 0:
return {'rRNA': str(rrna), 'rRNA_rate': "NA"}
rrna_rate = float(rrna) / sum(total_exp)
return {'rRNA': str(rrna), 'rRNA_rate': str(rrna_rate)}
def index(ref_file, out_dir, data):
"""Create a bismark index in the defined reference directory.
"""
(ref_dir, local_file) = os.path.split(ref_file)
gtf_file = dd.get_transcriptome_gtf(data, default=dd.get_gtf_file(data))
bismark = config_utils.find_program("bismark", data["config"])
if not utils.file_exists(gtf_file):
raise ValueError("%s not found, could not create a bismark index." % (gtf_file))
if not utils.file_exists(out_dir):
with tx_tmpdir(data, os.path.dirname(out_dir)) as tx_out_dir:
num_cores = dd.get_cores(data)
other_opts = config_utils.get_resources("bismark", data["config"]).get("options", [])
other_opts = " ".join([str(x) for x in other_opts]).strip()
cmd = "{bismark} {other_opts} --bowtie2 -p {num_cores} -n 1 -o {tx_out_dir} --basename {sample} --unmapped {ref_file} {in_fastq}"
do.run(cmd.format(**locals()), "Index STAR")
if os.path.exists(out_dir):
shutil.rmtree(out_dir)
shutil.move(tx_out_dir, out_dir)
return out_dir
def combine_express(samples, combined):
"""Combine tpm, effective counts and fpkm from express results"""
if not combined:
return None
to_combine = [
dd.get_express_counts(x) for x in dd.sample_data_iterator(samples)
if dd.get_express_counts(x)
]
gtf_file = dd.get_gtf_file(samples[0][0])
isoform_to_gene_file = os.path.join(os.path.dirname(combined),
"isoform_to_gene.txt")
isoform_to_gene_file = express.isoform_to_gene_name(
gtf_file, isoform_to_gene_file,
dd.sample_data_iterator(samples).next())
if len(to_combine) > 0:
eff_counts_combined_file = os.path.splitext(
combined)[0] + ".isoform.express_counts"
eff_counts_combined = count.combine_count_files(
to_combine, eff_counts_combined_file, ext=".counts")
to_combine = [
dd.get_express_tpm(x) for x in dd.sample_data_iterator(samples)
if dd.get_express_tpm(x)
]
tpm_counts_combined_file = os.path.splitext(
combined)[0] + ".isoform.express_tpm"
tpm_counts_combined = count.combine_count_files(
to_combine, tpm_counts_combined_file)
to_combine = [
dd.get_express_fpkm(x) for x in dd.sample_data_iterator(samples)
if dd.get_express_fpkm(x)
]
fpkm_counts_combined_file = os.path.splitext(
combined)[0] + ".isoform.express_fpkm"
fpkm_counts_combined = count.combine_count_files(
to_combine, fpkm_counts_combined_file, ext=".fpkm")
return {
'counts': eff_counts_combined,
'tpm': tpm_counts_combined,
'fpkm': fpkm_counts_combined,
'isoform_to_gene': isoform_to_gene_file
}
return {}
def sample_summary(bam_file, data, out_dir):
"""Run RNA-SeQC on a single RNAseq sample, writing to specified output directory.
"""
metrics_file = os.path.join(out_dir, "metrics.tsv")
if not file_exists(metrics_file):
with file_transaction(data, out_dir) as tx_out_dir:
config = data["config"]
ref_file = data["sam_ref"]
genome_dir = os.path.dirname(os.path.dirname(ref_file))
gtf_file = dd.get_gtf_file(data)
sample_file = os.path.join(safe_makedir(tx_out_dir), "sample_file.txt")
_write_sample_id_file(data, bam_file, sample_file)
runner = rnaseqc_runner_from_config(config)
rna_file = config_utils.get_rRNA_sequence(genome_dir)
bam.index(bam_file, config)
single_end = not bam.is_paired(bam_file)
runner.run(sample_file, ref_file, rna_file, gtf_file, tx_out_dir, single_end)
# we don't need this large directory for just the report
shutil.rmtree(os.path.join(tx_out_dir, data["description"]))
return _parse_rnaseqc_metrics(metrics_file, data["name"][-1])
def run_pizzly(data):
work_dir = dd.get_work_dir(data)
pizzlydir = os.path.join(work_dir, "pizzly")
samplename = dd.get_sample_name(data)
gtf = dd.get_gtf_file(data)
if dd.get_transcriptome_fasta(data):
gtf_fa = dd.get_transcriptome_fasta(data)
else:
gtf_fa = sailfish.create_combined_fasta(data)
stripped_fa = os.path.splitext(
os.path.basename(gtf_fa))[0] + "-noversions.fa"
stripped_fa = os.path.join(pizzlydir, stripped_fa)
gtf_fa = fasta.strip_transcript_versions(gtf_fa, stripped_fa)
fraglength = get_fragment_length(data)
cachefile = os.path.join(pizzlydir, "pizzly.cache")
fusions = kallisto.get_kallisto_fusions(data)
pizzlypath = config_utils.get_program("pizzly", dd.get_config(data))
outdir = pizzly(pizzlypath, gtf, gtf_fa, fraglength, cachefile, pizzlydir,
fusions, samplename, data)
return outdir
def _rnaseq_qualimap(bam_file, data, out_dir):
"""
Run qualimap for a rnaseq bam file and parse results
"""
report_file = os.path.join(out_dir, "qualimapReport.html")
config = data["config"]
gtf_file = dd.get_gtf_file(data)
ref_file = dd.get_ref_file(data)
single_end = not bam.is_paired(bam_file)
if not utils.file_exists(report_file):
utils.safe_makedir(out_dir)
bam.index(bam_file, config)
cmd = _rnaseq_qualimap_cmd(config, bam_file, out_dir, gtf_file, single_end)
do.run(cmd, "Qualimap for {}".format(data["name"][-1]))
metrics = _parse_rnaseq_qualimap_metrics(report_file)
metrics.update(_detect_duplicates(bam_file, out_dir, data))
metrics.update(_detect_rRNA(data))
metrics.update({"Fragment Length Mean": bam.estimate_fragment_size(bam_file)})
metrics = _parse_metrics(metrics)
return metrics
def run_salmon_reads(data):
data = utils.to_single_data(data)
files = dd.get_input_sequence_files(data)
if bam.is_bam(files[0]):
files = fastq.convert_bam_to_fastq(files[0], data["dirs"]["work"],
data, data["dirs"], data["config"])
samplename = dd.get_sample_name(data)
work_dir = dd.get_work_dir(data)
salmon_dir = os.path.join(work_dir, "salmon", samplename)
gtf_file = dd.get_gtf_file(data)
if len(files) == 2:
fq1, fq2 = files
else:
fq1, fq2 = files[0], None
fasta_file = dd.get_ref_file(data)
out_file = salmon_quant_reads(fq1, fq2, salmon_dir, gtf_file, fasta_file,
data)
data = dd.set_salmon(data, out_file)
data = dd.set_salmon_dir(data, salmon_dir)
return [[data]]
def run(data):
"""Quantitaive isoforms expression by eXpress"""
name = dd.get_sample_name(data)
in_bam = dd.get_transcriptome_bam(data)
config = data['config']
if not in_bam:
logger.info(
"Transcriptome-mapped BAM file not found, skipping eXpress.")
return data
out_dir = os.path.join(dd.get_work_dir(data), "express", name)
out_file = os.path.join(out_dir, name + ".xprs")
express = config_utils.get_program("express", data['config'])
strand = _set_stranded_flag(in_bam, data)
if not file_exists(out_file):
gtf_fasta = gtf.gtf_to_fasta(dd.get_gtf_file(data),
dd.get_ref_file(data))
with tx_tmpdir(data) as tmp_dir:
with file_transaction(data, out_dir) as tx_out_dir:
bam_file = _prepare_bam_file(in_bam, tmp_dir, config)
cmd = (
"{express} --no-update-check -o {tx_out_dir} {strand} {gtf_fasta} {bam_file}"
)
do.run(cmd.format(**locals()), "Run express on %s." % in_bam,
{})
shutil.move(os.path.join(out_dir, "results.xprs"), out_file)
eff_count_file = _get_column(out_file,
out_file.replace(".xprs", "_eff.counts"),
7,
data=data)
tpm_file = _get_column(out_file,
out_file.replace("xprs", "tpm"),
14,
data=data)
fpkm_file = _get_column(out_file,
out_file.replace("xprs", "fpkm"),
10,
data=data)
data = dd.set_express_counts(data, eff_count_file)
data = dd.set_express_tpm(data, tpm_file)
data = dd.set_express_fpkm(data, fpkm_file)
return data
def estimate_expression(samples, run_parallel):
samples = run_parallel("generate_transcript_counts", samples)
combined = count.combine_count_files(
[x[0]["count_file"] for x in samples if "count_file" in x[0]])
gtf_file = dd.get_gtf_file(samples[0][0], None)
annotated = count.annotate_combined_count_file(combined, gtf_file)
samples = run_parallel("run_cufflinks", samples)
#gene
fpkm_combined_file = os.path.splitext(combined)[0] + ".fpkm"
to_combine = [x[0]["fpkm"] for x in samples if "fpkm" in x[0]]
fpkm_combined = count.combine_count_files(to_combine, fpkm_combined_file)
#isoform
fpkm_isoform_combined_file = os.path.splitext(
combined)[0] + ".isoform.fpkm"
to_combine_isoform = [
x[0]["fpkm_isoform"] for x in samples if "fpkm_isoform" in x[0]
]
fpkm_isoform_combined = count.combine_count_files(
to_combine_isoform, fpkm_isoform_combined_file, ".isoform.fpkm")
dexseq_combined_file = os.path.splitext(combined)[0] + ".dexseq"
to_combine_dexseq = [dd.get_dexseq_counts(data[0]) for data in samples]
to_combine_dexseq = filter(lambda x: x, to_combine_dexseq)
if to_combine_dexseq:
dexseq_combined = count.combine_count_files(to_combine_dexseq,
dexseq_combined_file,
".dexseq")
else:
dexseq_combined = None
for x in samples:
x[0]["combined_counts"] = combined
if annotated:
x[0]["annotated_combined_counts"] = annotated
if fpkm_combined:
x[0]["combined_fpkm"] = fpkm_combined
if fpkm_isoform_combined:
x[0]["combined_fpkm_isoform"] = fpkm_isoform_combined
if dexseq_combined:
x[0] = dd.set_dexseq_counts(x[0], dexseq_combined_file)
return samples
def align_transcriptome(fastq_file, pair_file, ref_file, data):
"""
bowtie2 with settings for aligning to the transcriptome for eXpress/RSEM/etc
"""
work_bam = dd.get_work_bam(data)
base, ext = os.path.splitext(work_bam)
out_file = base + ".transcriptome" + ext
if file_exists(out_file):
data = dd.set_transcriptome_bam(data, out_file)
return data
bowtie2 = config_utils.get_program("bowtie2", data["config"])
gtf_file = dd.get_gtf_file(data)
gtf_index = index_transcriptome(gtf_file, ref_file, data)
num_cores = data["config"]["algorithm"].get("num_cores", 1)
pair_cmd = "-2 %s " % pair_file if pair_file else ""
cmd = ("{bowtie2} -p {num_cores} -a -X 600 --rdg 6,5 --rfg 6,5 --score-min L,-.6,-.4 --no-discordant --no-mixed -x {gtf_index} -1 {fastq_file} {pair_cmd} | samtools view -hbS - > {tx_out_file}")
with file_transaction(out_file) as tx_out_file:
message = "Aligning %s and %s to the transcriptome." % (fastq_file, pair_file)
do.run(cmd.format(**locals()), message)
data = dd.set_transcriptome_bam(data, out_file)
return data
def run_salmon_decoy(data):
data = utils.to_single_data(data)
files = dd.get_input_sequence_files(data)
if bam.is_bam(files[0]):
files = fastq.convert_bam_to_fastq(files[0], data["dirs"]["work"],
data, data["dirs"], data["config"])
samplename = dd.get_sample_name(data)
work_dir = dd.get_work_dir(data)
salmon_dir = os.path.join(work_dir, "salmon", samplename)
gtf_file = dd.get_gtf_file(data)
if len(files) == 2:
fq1, fq2 = files
else:
fq1, fq2 = files[0], None
index = salmon_decoy_index(gtf_file, data, os.path.dirname(salmon_dir))
out_file = salmon_quant_reads(fq1, fq2, salmon_dir, gtf_file, data, index)
data = dd.set_salmon(data, out_file)
data = dd.set_salmon_dir(data, salmon_dir)
data = dd.set_salmon_fraglen_file(data, _get_fraglen_file(salmon_dir))
data = dd.update_summary_qc(data, "salmon", base=dd.get_salmon_fraglen_file(data))
return [[data]]
def count(data):
"""
count reads mapping to genes using featureCounts
falls back on htseq_count method if featureCounts is not
found
"""
in_bam = dd.get_work_bam(data)
gtf_file = dd.get_gtf_file(data)
work_dir = dd.get_work_dir(data)
out_dir = os.path.join(work_dir, "htseq-count")
safe_makedir(out_dir)
count_file = os.path.join(out_dir, dd.get_sample_name(data)) + ".counts"
if file_exists(count_file):
return count_file
config = data["config"]
try:
featureCounts = config_utils.get_program("featureCounts", config)
except config_utils.CmdNotFound:
logger.info("featureCounts not found, falling back to htseq-count "
"for feature counting. You can upgrade the tools to "
"install featureCount with bcbio_nextgen.py upgrade "
"--tools.")
return htseq_count(data)
paired_flag = _paired_flag(in_bam)
strand_flag = _strand_flag(config)
cmd = ("{featureCounts} -a {gtf_file} -o {tx_count_file} -s {strand_flag} "
"{paired_flag} {in_bam}")
message = ("Count reads in {tx_count_file} mapping to {gtf_file} using "
"featureCounts")
with file_transaction(data, count_file) as tx_count_file:
do.run(cmd.format(**locals()), message.format(**locals()))
fixed_count_file = _format_count_file(count_file, data)
os.rename(fixed_count_file, count_file)
return count_file
def count(data):
"""
count reads mapping to genes using featureCounts
http://subread.sourceforge.net
"""
in_bam = dd.get_work_bam(data)
sorted_bam = bam.sort(in_bam, dd.get_config(data), order="queryname")
gtf_file = dd.get_gtf_file(data)
work_dir = dd.get_work_dir(data)
out_dir = os.path.join(work_dir, "htseq-count")
safe_makedir(out_dir)
count_file = os.path.join(out_dir, dd.get_sample_name(data)) + ".counts"
summary_file = os.path.join(out_dir,
dd.get_sample_name(data)) + ".counts.summary"
if file_exists(count_file):
return count_file
featureCounts = config_utils.get_program("featureCounts",
dd.get_config(data))
paired_flag = _paired_flag(in_bam)
strand_flag = _strand_flag(data)
filtered_bam = bam.filter_primary(sorted_bam, data)
cmd = ("{featureCounts} -a {gtf_file} -o {tx_count_file} -s {strand_flag} "
"{paired_flag} {filtered_bam}")
message = ("Count reads in {tx_count_file} mapping to {gtf_file} using "
"featureCounts")
with file_transaction(data, [count_file, summary_file]) as tx_files:
tx_count_file, tx_summary_file = tx_files
do.run(cmd.format(**locals()), message.format(**locals()))
fixed_count_file = _format_count_file(count_file, data)
fixed_summary_file = _change_sample_name(summary_file,
dd.get_sample_name(data),
data=data)
shutil.move(fixed_count_file, count_file)
shutil.move(fixed_summary_file, summary_file)
return count_file
def run_rnaseq(bam_file, data, out_dir):
"""
Run qualimap for a rnaseq bam file and parse results
"""
strandedness = {"firststrand": "strand-specific-reverse",
"secondstrand": "strand-specific-forward",
"unstranded": "non-strand-specific"}
# Qualimap results should be saved to a directory named after sample.
# MultiQC (for parsing additional data) picks the sample name after the dir as follows:
# <sample name>/raw_data_qualimapReport/insert_size_histogram.txt
results_dir = os.path.join(out_dir, dd.get_sample_name(data))
results_file = os.path.join(results_dir, "rnaseq_qc_results.txt")
report_file = os.path.join(results_dir, "qualimapReport.html")
config = data["config"]
gtf_file = dd.get_gtf_file(data)
single_end = not bam.is_paired(bam_file)
library = strandedness[dd.get_strandedness(data)]
if not utils.file_exists(results_file):
with file_transaction(data, results_dir) as tx_results_dir:
utils.safe_makedir(tx_results_dir)
bam.index(bam_file, config)
cmd = _rnaseq_qualimap_cmd(data, bam_file, tx_results_dir, gtf_file, single_end, library)
do.run(cmd, "Qualimap for {}".format(dd.get_sample_name(data)))
tx_results_file = os.path.join(tx_results_dir, "rnaseq_qc_results.txt")
cmd = "sed -i 's/bam file = .*/bam file = %s.bam/' %s" % (dd.get_sample_name(data), tx_results_file)
do.run(cmd, "Fix Name Qualimap for {}".format(dd.get_sample_name(data)))
metrics = _parse_rnaseq_qualimap_metrics(report_file)
metrics.update(_detect_duplicates(bam_file, results_dir, data))
metrics.update(_detect_rRNA(data, results_dir))
metrics.update({"Average_insert_size": bam.estimate_fragment_size(bam_file)})
metrics = _parse_metrics(metrics)
# Qualimap output folder (results_dir) needs to be named after the sample (see comments above). However, in order
# to keep its name after upload, we need to put the base QC file (results_file) into the root directory (out_dir):
base_results_file = os.path.join(out_dir, os.path.basename(results_file))
shutil.copyfile(results_file, base_results_file)
return {"base": base_results_file,
"secondary": _find_qualimap_secondary_files(results_dir, base_results_file),
"metrics": metrics}
def create_combined_tx2gene(data):
out_dir = os.path.join(dd.get_work_dir(data), "inputs", "transcriptome")
items = disambiguate.split([data])
tx2gene_files = []
for i in items:
odata = i[0]
gtf_file = dd.get_gtf_file(odata)
out_file = os.path.join(out_dir, dd.get_genome_build(odata) + "-tx2gene.csv")
if file_exists(out_file):
tx2gene_files.append(out_file)
else:
out_file = gtf.tx2genefile(gtf_file, out_file, tsv=False)
tx2gene_files.append(out_file)
combined_file = os.path.join(out_dir, "tx2gene.csv")
if file_exists(combined_file):
return combined_file
tx2gene_file_string = " ".join(tx2gene_files)
cmd = "cat {tx2gene_file_string} > {tx_out_file}"
with file_transaction(data, combined_file) as tx_out_file:
do.run(cmd.format(**locals()), "Combining tx2gene CSV files.")
return combined_file
def _stringtie_expression(bam, data, out_dir="."):
"""
only estimate expression the Stringtie, do not assemble new transcripts
"""
gtf_file = dd.get_transcriptome_gtf(data, dd.get_gtf_file(data))
num_cores = dd.get_num_cores(data)
error_message = "The %s file for %s is missing. StringTie has an error."
stringtie = config_utils.get_program("stringtie", data, default="stringtie")
# don't assemble transcripts unless asked
exp_flag = ("-e" if "stringtie" not in dd.get_transcript_assembler(data)
else "")
base_cmd = ("{stringtie} {exp_flag} -b {out_dir} -p {num_cores} -G {gtf_file} "
"-o {out_gtf} {bam}")
transcript_file = os.path.join(out_dir, "t_data.ctab")
exon_file = os.path.join(out_dir, "e_data.ctab")
out_gtf = os.path.join(out_dir, "stringtie-assembly.gtf")
if file_exists(transcript_file):
return exon_file, transcript_file, out_gtf
cmd = base_cmd.format(**locals())
do.run(cmd, "Running Stringtie on %s." % bam)
assert file_exists(exon_file), error_message % ("exon", exon_file)
assert file_exists(transcript_file), error_message % ("transcript", transcript_file)
return transcript_file
def run_rnaseq(bam_file, data, out_dir):
"""
Run qualimap for a rnaseq bam file and parse results
"""
strandedness = {
"firststrand": "strand-specific-reverse",
"secondstrand": "strand-specific-forward",
"unstranded": "non-strand-specific"
}
# Qualimap results should be saved to a directory named after sample.
# MultiQC (for parsing additional data) picks the sample name after the dir as follows:
# <sample name>/raw_data_qualimapReport/insert_size_histogram.txt
results_dir = os.path.join(out_dir, dd.get_sample_name(data))
report_file = os.path.join(results_dir, "qualimapReport.html")
raw_file = os.path.join(results_dir, "rnaseq_qc_results.txt")
config = data["config"]
gtf_file = dd.get_gtf_file(data)
single_end = not bam.is_paired(bam_file)
library = strandedness[dd.get_strandedness(data)]
if not utils.file_exists(report_file):
utils.safe_makedir(results_dir)
bam.index(bam_file, config)
cmd = _rnaseq_qualimap_cmd(data, bam_file, results_dir, gtf_file,
single_end, library)
do.run(cmd, "Qualimap for {}".format(dd.get_sample_name(data)))
cmd = "sed -i 's/bam file = .*/bam file = %s.bam/' %s" % (
dd.get_sample_name(data), raw_file)
do.run(cmd,
"Fix Name Qualimap for {}".format(dd.get_sample_name(data)))
metrics = _parse_rnaseq_qualimap_metrics(report_file)
metrics.update(_detect_duplicates(bam_file, results_dir, data))
metrics.update(_detect_rRNA(data))
metrics.update(
{"Average_insert_size": bam.estimate_fragment_size(bam_file)})
metrics = _parse_metrics(metrics)
return metrics
def rnaseq_vardict_variant_calling(data):
sample = dd.get_sample_name(data)
variation_dir = os.path.join(dd.get_work_dir(data), "variation")
safe_makedir(variation_dir)
out_file = os.path.join(variation_dir, sample + "-vardict.vcf.gz")
if file_exists(out_file):
data = dd.set_vrn_file(data, out_file)
return data
vardict_cmd = vardict.get_vardict_command(data)
strandbias = "teststrandbias.R"
var2vcf = "var2vcf_valid.pl"
vcfstreamsort = config_utils.get_program("vcfstreamsort", data)
compress_cmd = "| bgzip -c"
freq = float(dd.get_min_allele_fraction(data, 20) / 100.0)
var2vcf_opts = "-v 50"
fix_ambig = vcfutils.fix_ambiguous_cl()
remove_dup = vcfutils.remove_dup_cl()
r_setup = get_R_exports()
ref_file = dd.get_ref_file(data)
bamfile = dd.get_work_bam(data)
bed_file = gtf.gtf_to_bed(dd.get_gtf_file(data))
opts = " -c 1 -S 2 -E 3 -g 4 "
resources = config_utils.get_resources("vardict", data)
if resources.get("options"):
opts += " ".join([str(x) for x in resources["options"]])
with file_transaction(data, out_file) as tx_out_file:
jvm_opts = vardict._get_jvm_opts(data, tx_out_file)
cmd = ("{r_setup} && {jvm_opts}{vardict_cmd} -G {ref_file} -f {freq} "
"-N {sample} -b {bamfile} {opts} {bed_file} "
"| {strandbias}"
"| {var2vcf} -N {sample} -E -f {freq} {var2vcf_opts} "
"| {fix_ambig} | {remove_dup} | {vcfstreamsort} {compress_cmd} "
"> {tx_out_file}")
message = "Calling RNA-seq variants with VarDict"
do.run(cmd.format(**locals()), message)
data = dd.set_vrn_file(data, out_file)
return data
