def bamCoverageRNA(infile, outfile):
'''Make normalised bigwig tracks with deeptools'''
norm_method = PARAMS["deeptools_norm_method"]
# STAR MAPQ of 255 indicates uniquely mapped read
if len(infile) > 0:
if BamTools.is_paired(infile):
statement = f'''bamCoverage -b {infile} -o {outfile}
--binSize 5
--normalizeUsing {norm_method}
--samFlagInclude 64
--centerReads
--minMappingQuality 255
--smoothLength 10
--skipNAs
-p "max" '''
else:
statement = f'''bamCoverage -b {infile} -o {outfile}
--binSize 5
--normalizeUsing {norm_method}
--minMappingQuality 255
--smoothLength 10
--samFlagExclude 4
--centerReads
-p "max" '''
P.run(statement, job_memory="2G", job_threads=10)
def runRMATS(gtffile, designfile, pvalue, strand, outdir, permute=0):
'''Module to generate rMATS statment
Module offers the option to permute group name labels and
calculates readlength, which must be identical in all reads.
Arguments
---------
gtffile: string
path to :term:`gtf` file
designfile: string
path to design file
pvalue: string
threshold for FDR testing
strand: string
strandedness option: can be 'fr-unstranded', 'fr-firststrand',
or 'fr-secondstrand'
outdir: string
directory path for rMATS results
permute : 1 or 0
option to activate random shuffling of sample groups
'''
design = Expression.ExperimentalDesign(designfile)
if permute == 1:
permutelist = design.table.group.tolist()
random.shuffle(permutelist)
design.table.group = permutelist
group1 = ",".join(
["%s.bam" % x for x in design.getSamplesInGroup(design.groups[0])])
with open(outdir + "/b1.txt", "w") as f:
f.write(group1)
group2 = ",".join(
["%s.bam" % x for x in design.getSamplesInGroup(design.groups[1])])
with open(outdir + "/b2.txt", "w") as f:
f.write(group2)
readlength = BamTools.estimateTagSize(design.samples[0]+".bam")
statement = '''rMATS
--b1 %(outdir)s/b1.txt
--b2 %(outdir)s/b2.txt
--gtf <(gunzip -c %(gtffile)s)
--od %(outdir)s
--readLength %(readlength)s
--cstat %(pvalue)s
--libType %(strand)s
''' % locals()
# if Paired End Reads
if BamTools.is_paired(design.samples[0]+".bam"):
statement += '''-t paired''' % locals()
statement += '''
> %(outdir)s/%(designfile)s.log
'''
P.run(statement, job_condaenv="splicing")
def BAMtotalcounts(infile, outfile):
'''Count total reads in BAM for normalisation'''
if bamtools.is_paired(infile):
statement = f'''samtools view -f 2 {infile} | wc -l | awk 'BEGIN {{OFS="\\t"}} {{print $0/2}}' > {outfile}'''
# count only reads mapped in proper pairs
else:
statement = f'''samtools view -F 4 {infile} | wc -l | awk 'BEGIN {{OFS="\\t"}} {{print $0}}' > {outfile}'''
# exclude unmapped reads
P.run(statement)
def countDEXSeq(infiles, outfile):
'''create counts for DEXSeq
Counts bam reads agains exon features in flattened gtf.
The required python script is provided by DEXSeq
and uses HTSeqCounts.
Parameters
----------
infile[0]: string
:term:`bam` file input
infile[1]: string
:term:`gff` output from buildGff function
outfile : string
A :term:`txt` file containing results
DEXSeq_strandedness : string
:term:`PARAMS`. Specifies strandedness, options
are 'yes', 'no' and 'reverse'
'''
infile, gfffile = infiles
ps = PYTHONSCRIPTSDIR
if BamTools.is_paired(infile):
paired = "yes"
else:
paired = "no"
strandedness = PARAMS["DEXSeq_strandedness"]
statement = '''python %(ps)s/dexseq_count.py
-p %(paired)s
-s %(strandedness)s
-r pos
-f bam %(gfffile)s %(infile)s %(outfile)s'''
P.run(statement, job_condaenv="splicing")
def scoreIntervalsBAM(infiles, outfile):
'''Count reads in bed intervals'''
interval, bam = infiles
tmp_file = bam.replace(".merge.bam", ".tmp")
if bamtools.is_paired(bam):
# -p flag specifes only to count paired reads
options = "-p"
else:
options = " "
statement = f'''bedtools multicov
{options}
-q 10
-bams {bam}
-bed <(cut -f1-7 {interval} )
> {outfile} &&
sed -i '1i \contig\\tstart\\tend\\tpeak_id\\tpeak_score\\twidth\\tfeature\\ttotal'
{outfile}'''
P.run(statement)
def convertReadsToIntervals(bamfile,
bedfile,
filtering_quality=None,
filtering_dedup=None,
filtering_dedup_method='picard',
filtering_nonunique=False):
'''convert reads in *bamfile* to *intervals*.
This method converts read data into intervals for
counting based methods.
This method is not appropriate for RNA-Seq.
Optional steps include:
For paired end data, pairs are merged and optionally
filtered by insert size.
Arguments
---------
bamfile : string
Filename of input file in :term:`bam` format.
bedfile : string
Filename of output file in :term:`bed` format.
filtering_quality : int
If set, remove reads with a quality score below given threshold.
filtering_dedup : bool
If True, deduplicate data.
filtering_dedup_method : string
Deduplication method. Possible options are ``picard`` and
``samtools``.
filtering_nonunique : bool
If True, remove non-uniquely matching reads.
'''
track = P.snip(bedfile, ".bed.gz")
is_paired = BamTools.is_paired(bamfile)
current_file = bamfile
tmpdir = P.get_temp_filename()
statement = ["mkdir %(tmpdir)s"]
nfiles = 0
if filtering_quality > 0:
next_file = "%(tmpdir)s/bam_%(nfiles)i.bam" % locals()
statement.append('''samtools view
-q %(filtering_quality)i -b
%(current_file)s
2>> %%(bedfile)s.quality.log
> %(next_file)s ''' % locals())
nfiles += 1
current_file = next_file
if filtering_nonunique:
next_file = "%(tmpdir)s/bam_%(nfiles)i.bam" % locals()
statement.append('''cat %(current_file)s
| cgat bam2bam
--method=filter
--filter-method=unique,mapped
--log=%%(bedfile)s.nonunique.log
2> %%(bedfile)s.nonunique.err
> %(next_file)s ''' % locals())
nfiles += 1
current_file = next_file
if filtering_dedup is not None:
# Picard's MarkDuplicates requries an explicit bam file.
next_file = "%(tmpdir)s/bam_%(nfiles)i.bam" % locals()
if filtering_dedup_method == 'samtools':
statement.append('''samtools rmdup - - ''')
elif filtering_dedup_method == 'picard':
statement.append('''picard MarkDuplicates
INPUT=%(current_file)s
OUTPUT=%(next_file)s
ASSUME_SORTED=TRUE
METRICS_FILE=%(bedfile)s.duplicate_metrics
REMOVE_DUPLICATES=TRUE
VALIDATION_STRINGENCY=SILENT
>& %%(bedfile)s.markdup.log ''' % locals())
nfiles += 1
current_file = next_file
if is_paired:
statement.append('''cat %(current_file)s
| cgat bam2bed
--merge-pairs
--min-insert-size=%(filtering_min_insert_size)i
--max-insert-size=%(filtering_max_insert_size)i
--log=%(bedfile)s.bam2bed.log
-
2> %(bedfile)s.bam2bed.err
| cgat bed2bed
--method=sanitize-genome
--genome-file=%(genome_dir)s/%(genome)s
--log=%(bedfile)s.sanitize.log
2> %(bedfile)s.sanitize.err
| cut -f 1,2,3,4
| sort -k1,1 -k2,2n
| bgzip > %(bedfile)s''')
else:
statement.append('''cat %(current_file)s
| cgat bam2bed
--log=%(bedfile)s.bam2bed.log
-
2> %(bedfile)s.bam2bed.err
| cgat bed2bed
--method=sanitize-genome
--genome-file=%(genome_dir)s/%(genome)s
--log=%(bedfile)s.sanitize.log
2> %(bedfile)s.sanitize.err
| cut -f 1,2,3,4
| sort -k1,1 -k2,2n
| bgzip > %(bedfile)s''')
statement.append("tabix -p bed %(bedfile)s >& %(bedfile)s.tabix.log")
statement.append("rm -rf %(tmpdir)s")
statement = " ; ".join(statement)
P.run(statement, job_memory="8G")