def Pipeline_split():
split_fastq = sequence_(
map_(json_.Split), apply_(reduce_(['gz_output_dir'],
json_.Total_json)),
map_(json_.Format_json))
return sequence_(split_fastq, )
def Pipeline_split():
split_fastq = sequence_(
map_(json_.Split),
apply_(
reduce_(['gz_output_dir'],json_.Total_json)
),
map_(json_.Format_json)
)
return sequence_(
split_fastq,
)
def Pipeline():
testing = wga_settings['test']
target = wga_settings['target']
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval',range(1,23) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
align_to_reference = sequence_(
apply_(
reduce_(['sample_name', 'library', 'platform', 'platform_unit', 'chunk'], pipes.AlignAndCleanALN)
),
)
if target:
remove_dup = sequence_(
reduce_(['sample_name'], picard.MERGE_SAMS)
)
else:
remove_dup = sequence_(
reduce_(['sample_name'], picard.MarkDuplicates)
)
preprocess_alignment = sequence_(
map_(samtools.IndexBam),
apply_(
split_([intervals],gatk.RealignerTargetCreator) #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
),
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
apply_(
reduce_(['sample_name'], gatk.BQSRGatherer),
map_(gatk.ApplyBQSR) #TODO I add BQSRGatherer as a parent with a hack inside ApplyBQSR.cmd
)
)
call_variants = sequence_(
apply_(
#reduce_(['interval'],gatk.HaplotypeCaller,tag={'vcf':'HaplotypeCaller'}),
reduce_split_(['interval'], [glm], gatk.UnifiedGenotyper, tag={'vcf': 'UnifiedGenotyper'}),
combine=True
),
reduce_(['vcf'], gatk.CombineVariants, 'Combine Into Raw VCFs'),
split_([glm],gatk.VQSR),
)
return sequence_(
align_to_reference,
remove_dup,
preprocess_alignment,
call_variants
)
def Pipeline():
testing = wga_settings['test']
target = wga_settings['target']
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval', list(range(1, 23)) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
align_to_reference = sequence_(
apply_(
reduce_([
'sample_name', 'library', 'platform', 'platform_unit', 'chunk'
], pipes.AlignAndCleanALN)), )
if target:
remove_dup = sequence_(reduce_(['sample_name'], picard.MERGE_SAMS))
else:
remove_dup = sequence_(reduce_(['sample_name'], picard.MarkDuplicates))
preprocess_alignment = sequence_(
map_(samtools.IndexBam),
apply_(
split_(
[intervals], gatk.RealignerTargetCreator
) #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
),
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
apply_(
reduce_(['sample_name'], gatk.BQSRGatherer),
map_(
gatk.ApplyBQSR
) #TODO I add BQSRGatherer as a parent with a hack inside ApplyBQSR.cmd
))
call_variants = sequence_(
apply_(
#reduce_(['interval'],gatk.HaplotypeCaller,tag={'vcf':'HaplotypeCaller'}),
reduce_split_(['interval'], [glm],
gatk.UnifiedGenotyper,
tag={'vcf': 'UnifiedGenotyper'}),
combine=True),
reduce_(['vcf'], gatk.CombineVariants, 'Combine Into Raw VCFs'),
split_([glm], gatk.VQSR),
)
return sequence_(align_to_reference, remove_dup, preprocess_alignment,
call_variants)
def Pipeline_upload():
download_fastq = sequence_(
map_(s3_.S3Upload)
)
return sequence_(
download_fastq,
)
def Pipeline_Somatic():
testing = wga_settings['test']
target = wga_settings['target']
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval',range(1,23) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
align_to_reference = sequence_(
apply_(
reduce_(['sample_name', 'library', 'platform', 'platform_unit','sample_type','chunk','rgid'], pipes.AlignAndCleanMEM)
),
)
if target:
remove_dup = sequence_(
reduce_(['sample_name','sample_type','rgid'], picard.MERGE_SAMS)
)
else:
remove_dup = sequence_(
reduce_(['sample_name','sample_type','rgid'], picard.MarkDuplicates)
)
preprocess_alignment = sequence_(
map_(samtools.IndexBam),
apply_(
split_([intervals],gatk.RealignerTargetCreator) #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
),
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
apply_(
reduce_(['sample_name','sample_type','rgid'], gatk.BQSRGatherer),
map_(gatk.ApplyBQSR) #TODO I add BQSRGatherer as a parent with a hack inside ApplyBQSR.cmd
)
)
somatic_call = sequence_(
apply_(
sequence_(
map_(mutect.createInput),
reduce_(['rgid','interval'], mutect.Somatic, tag={'vcf': 'Mutect'}),
reduce_(['vcf'], gatk.CombineVariants, 'Combine Into Raw VCFs'),
),
sequence_(
map_(svdetect.PreProcessing),
map_(svdetect.link2SV)
)
)
)
return sequence_(
align_to_reference,
remove_dup,
preprocess_alignment,
somatic_call
)
def gunzip(workflow,input_dir,**kwargs):
"""
Gunzips all gz files in directory
$ genomekey gunzip -n 'Gunzip' /path/to/dir
"""
DAG().sequence_(
add_([ INPUT(f,tags={'i':i}) for i,f in enumerate(glob.glob(os.path.join(input_dir,'*.gz'))) ]),
map_(unix.Gunzip),
add_run(workflow)
)
def gunzip(workflow, input_dir, **kwargs):
"""
Gunzips all gz files in directory
$ genomekey gunzip -n 'Gunzip' /path/to/dir
"""
DAG().sequence_(
add_([
INPUT(f, tags={'i': i})
for i, f in enumerate(glob.glob(os.path.join(input_dir, '*.gz')))
]), map_(unix.Gunzip), add_run(workflow))
def Pipeline_Somatic():
testing = wga_settings['test']
target = wga_settings['target']
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval', list(range(1, 23)) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
align_to_reference = sequence_(
apply_(
reduce_([
'sample_name', 'library', 'platform', 'platform_unit',
'sample_type', 'chunk', 'rgid'
], pipes.AlignAndCleanMEM)), )
if target:
remove_dup = sequence_(
reduce_(['sample_name', 'sample_type', 'rgid'], picard.MERGE_SAMS))
else:
remove_dup = sequence_(
reduce_(['sample_name', 'sample_type', 'rgid'],
picard.MarkDuplicates))
preprocess_alignment = sequence_(
map_(samtools.IndexBam),
apply_(
split_(
[intervals], gatk.RealignerTargetCreator
) #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
),
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
apply_(
reduce_(['sample_name', 'sample_type', 'rgid'], gatk.BQSRGatherer),
map_(
gatk.ApplyBQSR
) #TODO I add BQSRGatherer as a parent with a hack inside ApplyBQSR.cmd
))
somatic_call = sequence_(
apply_(
sequence_(
map_(mutect.createInput),
reduce_(['rgid', 'interval'],
mutect.Somatic,
tag={'vcf': 'Mutect'}),
reduce_(['vcf'], gatk.CombineVariants,
'Combine Into Raw VCFs'),
), sequence_(map_(svdetect.PreProcessing),
map_(svdetect.link2SV))))
return sequence_(align_to_reference, remove_dup, preprocess_alignment,
somatic_call)
def CteamPipeline(input_bams):
bam_seq = None
bam_dup = [] # to check duplicate input files
for b in input_bams:
# extract genome_id from file, add as a tag
genome_id = os.path.basename(b).partition('.')[0]
if genome_id in bam_dup:
print '\n\nERROR: \"%s\" was already included in the input file list.\n' % b
sys.exit()
else:
bam_dup.append(genome_id)
# TEMPORARILLY, use genome_id as RG_ID, too
s = sequence_(
add_([INPUT(b, tags={'rg': genome_id})], stage_name="Load Input"))
# append to sequence
if bam_seq is None: bam_seq = s
else: bam_seq = sequence_(bam_seq, s, combine=True)
nInput = len(input_bams)
nNodes = settings.settings['nNode']
nSplit = min(
256, 16 *
max(nNodes / nInput, 1)) # will use floor, min 16, up to 256 splits
settings.settings['nSplit'] = nSplit
chrom = ('chrom', range(1, 23) + ['X', 'Y', 'MT'])
split = ('split', range(1, nSplit + 1))
return sequence_(
bam_seq,
map_(pipes.CteamSortSplitBam), # sort bam by readname (== shuffling)
split_([split], pipes.CteamTrimReadGroup), #
map_(pipes.CteamBwaAln), # bwa aln
map_(pipes.CteamBwaSampe), # bwa sampe
reduce_(['rg'], pipes.CteamSplitByChromosome
), # merge split files and (re)split by chromosome
split_([chrom], pipes.CteamRmDup_BuildIndex), # samtools rmdup + index
map_(pipes.CteamRealignTarget), # gatk indel realign target creator
map_(pipes.CteamIndelRealigner), # gatk indel realigner
map_(pipes.CteamUnifiedGenotyper) # gatk unifiedGenotyper
# #map_(pipes.CteamVariantFiltration) # gatk variantFilter
)
def Pipeline_local():
testing = wga_settings['test']
target = wga_settings['target']
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval',range(1,23) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
print [glm][0]
align_to_reference = sequence_(
apply_(
reduce_(['sample_name', 'library', 'platform', 'platform_unit'], pipes.AlignAndCleanALN)
),
)
if target:
remove_dup = sequence_(
)
else:
remove_dup = sequence_(
reduce_(['sample_name'], picard.MarkDuplicates)
)
preprocess_alignment = sequence_(
map_(samtools.IndexBam),
map_(gatk.RealignerTargetCreator), #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
map_(gatk.ApplyBQSR_local)
)
call_variants = sequence_(
apply_(
map_(gatk.UnifiedGenotyper_local, tag={'vcf': 'UnifiedGenotyper'})
)
)
return sequence_(
align_to_reference,
remove_dup,
preprocess_alignment,
call_variants
)
def Pipeline_local():
testing = wga_settings['test']
target = wga_settings['target']
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval', list(range(1, 23)) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
print([glm][0])
align_to_reference = sequence_(
apply_(
reduce_(['sample_name', 'library', 'platform', 'platform_unit'],
pipes.AlignAndCleanALN)), )
if target:
remove_dup = sequence_()
else:
remove_dup = sequence_(reduce_(['sample_name'], picard.MarkDuplicates))
preprocess_alignment = sequence_(
map_(samtools.IndexBam),
map_(
gatk.RealignerTargetCreator
), #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
map_(gatk.ApplyBQSR_local))
call_variants = sequence_(
apply_(
map_(gatk.UnifiedGenotyper_local, tag={'vcf':
'UnifiedGenotyper'})))
return sequence_(align_to_reference, remove_dup, preprocess_alignment,
call_variants)
def Pipeline():
is_capture = wga_settings['capture']
testing = wga_settings['test']
# split_ tuples
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval',range(1,23) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
align_to_reference = sequence_(
apply_(
reduce_(['sample_name', 'library'], misc.FastqStats),
reduce_(['sample_name', 'library', 'platform', 'platform_unit', 'chunk'], pipes.AlignAndClean)
),
)
preprocess_alignment = sequence_(
reduce_(['sample_name'], picard.MarkDuplicates),
apply_(
map_(picard.CollectMultipleMetrics),
split_([intervals],gatk.RealignerTargetCreator) #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
),
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
apply_(
reduce_(['sample_name'], gatk.BQSRGatherer),
map_(gatk.ApplyBQSR) #TODO I add BQSRGatherer as a parent with a hack inside ApplyBQSR.cmd
)
)
call_variants = sequence_(
# apply_(
# reduce_split_([],[intervals,glm], gatk.UnifiedGenotyper, tag={'vcf': 'UnifiedGenotyper'}),
# combine=True
# ) if is_capture
# else
apply_(
#reduce_(['interval'],gatk.HaplotypeCaller,tag={'vcf':'HaplotypeCaller'}),
reduce_split_(['interval'], [glm], gatk.UnifiedGenotyper, tag={'vcf': 'UnifiedGenotyper'}),
combine=True
),
reduce_(['vcf'], gatk.CombineVariants, 'Combine Into Raw VCFs'),
split_([glm],gatk.VQSR),
map_(gatk.Apply_VQSR),
reduce_(['vcf'], gatk.CombineVariants, "Combine into Master VCFs")
)
if is_capture:
return sequence_(
align_to_reference,
preprocess_alignment,
call_variants,
massive_annotation
)
else:
return sequence_(
align_to_reference,
preprocess_alignment,
reduce_split_(['sample_name'],[intervals],gatk.ReduceReads),
call_variants,
massive_annotation
)
from cosmos.lib.ezflow.dag import add_,map_,reduce_,split_,reduce_split_,sequence_,branch_
from subprocess import Popen,PIPE
from genomekey.tools import annovarext
from genomekey.wga_settings import wga_settings
import sys
import os
def get_db_names():
cmd = '{0} listdbs'.format(wga_settings['annovarext_path'])
if not os.path.exists(wga_settings['annovarext_path']):
raise Exception, 'AnnovarExtensions is not installed at {0}'.format(wga_settings['annovarext_path'])
dbs = Popen(cmd.split(' '),stdout=PIPE).communicate()[0]
if len(dbs) < 10:
raise Exception, "could not list databases, command was {0}".format(cmd)
return [ db for db in dbs.split('\n') if db != '' ]
massive_annotation = sequence_(
map_(annovarext.Vcf2Anno_in),
split_( [('build',['hg19']),('dbname',get_db_names()) ], annovarext.Annotate ),
reduce_(['vcf'],annovarext.MergeAnnotations)
)
def pipeline(bams, test_bam=False, chromosome_only_split=False):
# split_ tuples
#chrom = ('chrom', range(1,23) + ['X', 'Y', 'MT'])
chrom = ('chrom', range(1,23))
glm = ('glm', ['SNP', 'INDEL'])
dbnames = ('dbname', ['dbSNP135','CytoBand','Target_Scan','mirBase','Self_Chain','Repeat_Masker','TFBS','Segmental_Duplications','SIFT','COSMIC',
'PolyPhen2','Mutation_Taster','GERP','PhyloP','LRT','Mce46way','Complete_Genomics_69','The_1000g_Febuary_all','The_1000g_April_all',
'NHLBI_Exome_Project_euro','NHLBI_Exome_Project_aa','NHLBI_Exome_Project_all','ENCODE_DNaseI_Hypersensitivity','ENCODE_Transcription_Factor',
'UCSC_Gene','Refseq_Gene','Ensembl_Gene','CCDS_Gene','HGMD_INDEL','HGMD_SNP','GWAS_Catalog'])
bam_seq = None
for b in bams:
header = _getHeaderInfo(b)
sn = _getSeqName(header)
rgid = [ h[0] for h in header['rg']]
# restrict output for testing
if test_bam:
sn = ['chr1']
chrom = ('chrom',[1])
glm = ('glm',['SNP'])
skip_VQSR = ('skip_VQSR', [True])
else:
skip_VQSR = ('skip_VQSR', [False])
# if seqName is empty, then let's assume that the input is unaligned bam
# use everything before extension as part of tag
sample_name = os.path.splitext(os.path.basename(b))[0]
if chromosome_only_split:
# Stop splitting by rgId
bam_bwa_split = [ ('prevSn', sn), ('chromosome_only_split', [True]) ]
indelrealign_reduce = ['bam']
else:
bam_bwa_split = [ ('rgId', rgid), ('prevSn', sn), ('chromosome_only_split', [False]) ]
indelrealign_reduce = ['bam','rgId']
s = sequence_( add_([INPUT(b, tags={'bam':sample_name})], stage_name="Load BAMs"),
split_(bam_bwa_split, pipes.Bam_To_BWA))
if bam_seq is None: bam_seq = s
else: bam_seq = sequence_(bam_seq, s, combine=True)
# Previous pipeline
pr_pipeline = sequence_(
bam_seq,
reduce_split_(indelrealign_reduce, [chrom], pipes.IndelRealigner),
map_( pipes.MarkDuplicates),
reduce_(['bam','chrom'], pipes.BaseQualityScoreRecalibration),
map_( pipes.ReduceReads),
reduce_split_(['chrom'], [glm], pipes.UnifiedGenotyper),
reduce_(['glm'], pipes.VariantQualityScoreRecalibration, tag={'vcf':'main'}),
reduce_(['vcf'], pipes.CombineVariants, "Merge VCF"),
map_( pipes.Vcf2Anno_in),
split_([dbnames], pipes.Annotate, tag={'build':'hg19'}),
reduce_(['vcf'], pipes.MergeAnnotations)
)
# HaplotypeCaller Pipeline: official for GATK 3.0
hc_pipeline = sequence_(
bam_seq,
reduce_split_(indelrealign_reduce, [chrom], pipes.IndelRealigner),
map_( pipes.MarkDuplicates),
reduce_(['bam','chrom'], pipes.BaseQualityScoreRecalibration),
map_( pipes.HaplotypeCaller),
reduce_(['chrom'], pipes.GenotypeGVCFs),
split_([glm, skip_VQSR], pipes.VariantQualityScoreRecalibration, tag={'vcf':'main'})
)
return hc_pipeline
from cosmos.Workflow.models import Workflow
from cosmos.lib.ezflow.dag import DAG, split_,add_,map_,reduce_
from tools import ECHO, CAT, WC, PASTE, Sleep
####################
# Workflow
####################
dag = DAG().sequence_(
add_([ ECHO(tags={'word':'hello'}), ECHO(tags={'word':'world'}) ]),
map_(Sleep),
split_([('i',[1,2])], CAT),
reduce_([], PASTE),
map_(WC),
)
dag.create_dag_img('/tmp/ex.svg')
#################
# Run Workflow
#################
WF = Workflow.start('Example 3',restart=True,delete_intermediates=True)
dag.add_to_workflow(WF)
WF.run()
from cosmos.Workflow.models import Workflow
from cosmos.lib.ezflow.dag import DAG, split_,add_,map_,reduce_
from tools import ECHO, CAT, WC, PASTE
####################
# Workflow
####################
dag = DAG().sequence_(
add_([ ECHO(tags={'word':'hello'}), ECHO(tags={'word':'world'}) ]),
split_([('i',[1,2])], CAT),
reduce_([], PASTE),
map_(WC)
)
dag.create_dag_img('/tmp/ex.svg')
#################
# Run Workflow
#################
WF = Workflow.start('Example 2',restart=True,delete_intermediates=True)
dag.add_to_workflow(WF)
WF.run()
from cosmos.lib.ezflow.dag import add_,map_,reduce_,split_,reduce_split_,sequence_,branch_
from subprocess import Popen,PIPE
from genomekey.tools import annovarext
from genomekey.wga_settings import wga_settings
import sys
import os
def get_db_names():
cmd = '{0} listdbs'.format(wga_settings['annovarext_path'])
if not os.path.exists(wga_settings['annovarext_path']):
raise Exception('AnnovarExtensions is not installed at {0}'.format(wga_settings['annovarext_path']))
dbs = Popen(cmd.split(' '),stdout=PIPE).communicate()[0]
if len(dbs) < 10:
raise Exception("could not list databases, command was {0}".format(cmd))
return [ db for db in dbs.split('\n') if db != '' ]
massive_annotation = sequence_(
map_(annovarext.Vcf2Anno_in),
split_( [('build',['hg19']),('dbname',get_db_names()) ], annovarext.Annotate ),
reduce_(['vcf'],annovarext.MergeAnnotations)
)
def Pipeline():
is_capture = wga_settings['capture']
testing = wga_settings['test']
# split_ tuples
if testing:
intervals = ('interval', [20])
else:
intervals = ('interval', range(1, 23) + ['X', 'Y'])
glm = ('glm', ['SNP', 'INDEL'])
align_to_reference = sequence_(
apply_(
reduce_(['sample_name', 'library'], misc.FastqStats),
reduce_([
'sample_name', 'library', 'platform', 'platform_unit', 'chunk'
], pipes.AlignAndClean)), )
preprocess_alignment = sequence_(
reduce_(['sample_name'], picard.MarkDuplicates),
apply_(
map_(picard.CollectMultipleMetrics),
split_(
[intervals], gatk.RealignerTargetCreator
) #if not is_capture or testing else map_(gatk.RealignerTargetCreator)
),
map_(gatk.IndelRealigner),
map_(gatk.BQSR),
apply_(
reduce_(['sample_name'], gatk.BQSRGatherer),
map_(
gatk.ApplyBQSR
) #TODO I add BQSRGatherer as a parent with a hack inside ApplyBQSR.cmd
))
call_variants = sequence_(
# apply_(
# reduce_split_([],[intervals,glm], gatk.UnifiedGenotyper, tag={'vcf': 'UnifiedGenotyper'}),
# combine=True
# ) if is_capture
# else
apply_(
#reduce_(['interval'],gatk.HaplotypeCaller,tag={'vcf':'HaplotypeCaller'}),
reduce_split_(['interval'], [glm],
gatk.UnifiedGenotyper,
tag={'vcf': 'UnifiedGenotyper'}),
combine=True),
reduce_(['vcf'], gatk.CombineVariants, 'Combine Into Raw VCFs'),
split_([glm], gatk.VQSR),
map_(gatk.Apply_VQSR),
reduce_(['vcf'], gatk.CombineVariants, "Combine into Master VCFs"))
if is_capture:
return sequence_(align_to_reference, preprocess_alignment,
call_variants, massive_annotation)
else:
return sequence_(
align_to_reference, preprocess_alignment,
reduce_split_(['sample_name'], [intervals], gatk.ReduceReads),
call_variants, massive_annotation)
def Pipeline_upload():
download_fastq = sequence_(map_(s3_.S3Upload))
return sequence_(download_fastq, )
from cosmos.Workflow.models import Workflow
from cosmos.lib.ezflow.dag import DAG, split_, add_, map_, reduce_
from tools import ECHO, CAT, WC, PASTE
####################
# Workflow
####################
dag = DAG().sequence_(
add_([ECHO(tags={'word': 'hello'}),
ECHO(tags={'word': 'world'})]), split_([('i', [1, 2])], CAT),
reduce_([], PASTE), map_(WC))
dag.create_dag_img('/tmp/ex.svg')
#################
# Run Workflow
#################
WF = Workflow.start('Example 2', restart=True, delete_intermediates=True)
dag.add_to_workflow(WF)
WF.run()
def pipeline(bams, test_bam=False, chromosome_only_split=False):
# split_ tuples
#chrom = ('chrom', range(1,23) + ['X', 'Y', 'MT'])
chrom = ('chrom', range(1, 23))
glm = ('glm', ['SNP', 'INDEL'])
dbnames = ('dbname', [
'dbSNP135', 'CytoBand', 'Target_Scan', 'mirBase', 'Self_Chain',
'Repeat_Masker', 'TFBS', 'Segmental_Duplications', 'SIFT', 'COSMIC',
'PolyPhen2', 'Mutation_Taster', 'GERP', 'PhyloP', 'LRT', 'Mce46way',
'Complete_Genomics_69', 'The_1000g_Febuary_all', 'The_1000g_April_all',
'NHLBI_Exome_Project_euro', 'NHLBI_Exome_Project_aa',
'NHLBI_Exome_Project_all', 'ENCODE_DNaseI_Hypersensitivity',
'ENCODE_Transcription_Factor', 'UCSC_Gene', 'Refseq_Gene',
'Ensembl_Gene', 'CCDS_Gene', 'HGMD_INDEL', 'HGMD_SNP', 'GWAS_Catalog'
])
bam_seq = None
for b in bams:
header = _getHeaderInfo(b)
sn = _getSeqName(header)
rgid = [h[0] for h in header['rg']]
# restrict output for testing
if test_bam:
sn = ['chr1']
chrom = ('chrom', [1])
glm = ('glm', ['SNP'])
skip_VQSR = ('skip_VQSR', [True])
else:
skip_VQSR = ('skip_VQSR', [False])
# if seqName is empty, then let's assume that the input is unaligned bam
# use everything before extension as part of tag
sample_name = os.path.splitext(os.path.basename(b))[0]
if chromosome_only_split:
# Stop splitting by rgId
bam_bwa_split = [('prevSn', sn), ('chromosome_only_split', [True])]
indelrealign_reduce = ['bam']
else:
bam_bwa_split = [('rgId', rgid), ('prevSn', sn),
('chromosome_only_split', [False])]
indelrealign_reduce = ['bam', 'rgId']
s = sequence_(
add_([INPUT(b, tags={'bam': sample_name})],
stage_name="Load BAMs"),
split_(bam_bwa_split, pipes.Bam_To_BWA))
if bam_seq is None: bam_seq = s
else: bam_seq = sequence_(bam_seq, s, combine=True)
# Previous pipeline
pr_pipeline = sequence_(
bam_seq,
reduce_split_(indelrealign_reduce, [chrom], pipes.IndelRealigner),
map_(pipes.MarkDuplicates),
reduce_(['bam', 'chrom'], pipes.BaseQualityScoreRecalibration),
map_(pipes.ReduceReads),
reduce_split_(['chrom'], [glm], pipes.UnifiedGenotyper),
reduce_(['glm'],
pipes.VariantQualityScoreRecalibration,
tag={'vcf': 'main'}),
reduce_(['vcf'], pipes.CombineVariants, "Merge VCF"),
map_(pipes.Vcf2Anno_in),
split_([dbnames], pipes.Annotate, tag={'build': 'hg19'}),
reduce_(['vcf'], pipes.MergeAnnotations))
# HaplotypeCaller Pipeline: official for GATK 3.0
hc_pipeline = sequence_(
bam_seq,
reduce_split_(indelrealign_reduce, [chrom], pipes.IndelRealigner),
map_(pipes.MarkDuplicates),
reduce_(['bam', 'chrom'], pipes.BaseQualityScoreRecalibration),
map_(pipes.HaplotypeCaller), reduce_(['chrom'], pipes.GenotypeGVCFs),
split_([glm, skip_VQSR],
pipes.VariantQualityScoreRecalibration,
tag={'vcf': 'main'}))
return hc_pipeline
