def mark_dup_picard(bam, out_dir, java_heap=None): # shared by both se and pe
prefix = os.path.join(out_dir, os.path.basename(strip_ext_bam(bam)))
# strip extension appended in the previous step
prefix = strip_ext(prefix, 'filt')
dupmark_bam = '{}.dupmark.bam'.format(prefix)
dup_qc = '{}.dup.qc'.format(prefix)
if java_heap is None:
java_heap_param = '-Xmx4G'
else:
java_heap_param = '-Xmx{}'.format(java_heap)
run_shell_cmd('java {java_heap_param} -XX:ParallelGCThreads=1 '
'-jar {picard} MarkDuplicates '
'INPUT={bam} '
'OUTPUT={dupmark_bam} '
'METRICS_FILE={dup_qc} '
'VALIDATION_STRINGENCY=LENIENT '
'USE_JDK_DEFLATER=TRUE '
'USE_JDK_INFLATER=TRUE '
'ASSUME_SORTED=TRUE '
'REMOVE_DUPLICATES=FALSE '.format(
java_heap_param=java_heap_param,
picard=locate_picard(),
bam=bam,
dupmark_bam=dupmark_bam,
dup_qc=dup_qc,
))
return dupmark_bam, dup_qc
def get_gc(qsorted_bam_file, reference_fasta, prefix, java_heap=None):
'''
Uses picard tools (CollectGcBiasMetrics). Note that the reference
MUST be the same fasta file that generated the bowtie indices.
Assumes picard was already loaded into space (module add picard-tools)
'''
# remove redundant (or malformed) info (read group) from bam
logging.info('Getting GC bias...')
output_file = '{0}.gc.txt'.format(prefix)
plot_file = '{0}.gcPlot.pdf'.format(prefix)
summary_file = '{0}.gcSummary.txt'.format(prefix)
if java_heap is None:
java_heap_param = '-Xmx10G'
else:
java_heap_param = '-Xmx{}'.format(java_heap)
get_gc_metrics = ('java {6} -XX:ParallelGCThreads=1 -jar '
'{5} '
'CollectGcBiasMetrics R={0} I={1} O={2} '
'USE_JDK_DEFLATER=TRUE USE_JDK_INFLATER=TRUE '
'VERBOSITY=ERROR QUIET=TRUE '
'ASSUME_SORTED=FALSE '
'CHART={3} S={4}').format(reference_fasta,
qsorted_bam_file,
output_file,
plot_file,
summary_file,
locate_picard(),
java_heap_param)
logging.info(get_gc_metrics)
os.system(get_gc_metrics)
return output_file, plot_file, summary_file
def get_picard_complexity_metrics(aligned_bam, prefix, java_heap=None):
'''
Picard EsimateLibraryComplexity
'''
# remove redundant (or malformed) info (read group) from bam
out_file = '{0}.picardcomplexity.qc'.format(prefix)
if java_heap is None:
java_heap_param = '-Xmx6G'
else:
java_heap_param = '-Xmx{}'.format(java_heap)
get_gc_metrics = (
'mkdir -p tmp_java && java -Djava.io.tmpdir=$PWD/tmp_java '
'{3} -XX:ParallelGCThreads=1 -jar '
'{2} '
'EstimateLibraryComplexity INPUT={0} OUTPUT={1} '
'USE_JDK_DEFLATER=TRUE USE_JDK_INFLATER=TRUE '
'VERBOSITY=ERROR '
'QUIET=TRUE && rm -rf tmp_java').format(aligned_bam, out_file,
locate_picard(),
java_heap_param)
os.system(get_gc_metrics)
# Extract the actual estimated library size
header_seen = False
est_library_size = 0
with open(out_file, 'r') as fp:
for line in fp:
if header_seen:
est_library_size = int(float(line.strip().split()[-1]))
break
if 'ESTIMATED_LIBRARY_SIZE' in line:
header_seen = True
return est_library_size
def get_insert_distribution(final_bam, prefix):
'''
Calls Picard CollectInsertSizeMetrics
'''
log.info('insert size distribution...')
insert_data = '{0}.inserts.hist_data.log'.format(prefix)
insert_plot = '{0}.inserts.hist_graph.pdf'.format(prefix)
graph_insert_dist = ('java -Xmx6G -XX:ParallelGCThreads=1 -jar '
'{3} '
'CollectInsertSizeMetrics '
'INPUT={0} OUTPUT={1} H={2} '
'VERBOSITY=ERROR QUIET=TRUE '
'USE_JDK_DEFLATER=TRUE USE_JDK_INFLATER=TRUE '
'W=1000 STOP_AFTER=5000000').format(
final_bam, insert_data, insert_plot,
locate_picard())
log.info(graph_insert_dist)
os.system(graph_insert_dist)
return insert_data, insert_plot
def mark_dup_picard(bam, out_dir): # shared by both se and pe
prefix = os.path.join(out_dir,
os.path.basename(strip_ext_bam(bam)))
# strip extension appended in the previous step
prefix = strip_ext(prefix, 'filt')
dupmark_bam = '{}.dupmark.bam'.format(prefix)
dup_qc = '{}.dup.qc'.format(prefix)
cmd = 'java -Xmx4G -XX:ParallelGCThreads=1 -jar '
cmd += locate_picard()
cmd += ' MarkDuplicates '
# cmd = 'picard MarkDuplicates '
cmd += 'INPUT={} OUTPUT={} '
cmd += 'METRICS_FILE={} VALIDATION_STRINGENCY=LENIENT '
cmd += 'USE_JDK_DEFLATER=TRUE USE_JDK_INFLATER=TRUE '
cmd += 'ASSUME_SORTED=true REMOVE_DUPLICATES=false'
cmd = cmd.format(
bam,
dupmark_bam,
dup_qc)
run_shell_cmd(cmd)
return dupmark_bam, dup_qc
