首先，使用ALLbio软件下的breakdancer2vcf.py转化breakdancer得到的out文件
GitHub地址：这里
代码如下：

 #!/usr/bin/env python

    __copyright__ = """
    Copyright (C) 2013 - Tim te Beek
    Copyright (C) 2013 - Wai Yi Leung
    Copyright (C) 2013 AllBio (see AUTHORS file)
    """

    __desc__ = """Convert breakdancer output to pseudo .vcf file format."""
    __created__ = "Mar 18, 2013"
    __author__ = "tbeek"
  
    import argparse
    import csv
    import os.path
    import sys


    def main(tsvfile, vcffile):
'''
:param tsvfile: filename of input file.tsv
:type tsvfile: string
:param vcffile: filename of output file.vcf
:type vcffile: string
'''
with open(tsvfile) as reader:
    # Parse file
    dictreader = _parse_tsvfile(reader)
    print dictreader.fieldnames

    # Write out file
    _format_vcffile(dictreader, vcffile)

# Quick output
with open(vcffile) as reader:
    print reader.read(1000)
    def _parse_tsvfile(readable):
'''
Read readable using csv.Sniffer and csv.DictReader
:param readable: open file.tsv handle to read with csv.DictReader
:type readable: file
'''
prev, curr = 0, 0
while True:
    line = readable.readline()
    if not line.startswith('#'):
        # lets start from prev # line, without the hash sign
        readable.seek(prev + 1)
        break
    else:
        prev = curr
        curr = readable.tell()

# Determine dialect
curr = readable.tell()
#dialect = csv.Sniffer().sniff(readable.read(3000))
dialect = 'excel-tab'
readable.seek(curr)

# Read file
dictreader = csv.DictReader(readable, dialect=dialect)
return dictreader


    _tsv_fields = ('Chr1', 'Pos1', 'Orientation1',
           'Chr2', 'Pos2', 'Orientation2',
           'Type', 'Size', 'Score',
           'num_Reads', 'num_Reads_lib',
           'ERR031544.sort.bam')
    # 'Chr1': '1',
    # 'Pos1': '269907',
    # 'Orientation1': '39+39-',
    # 'Chr2': '1',
    # 'Pos2': '270919',
    # 'Orientation2': '39+39-',
    # 'Type': 'DEL',
    # 'Size': '99',
    # 'Score': '99',
    # 'num_Reads': '38',
    # 'num_Reads_lib': '/home/allbio/ERR031544.sort.bam|38',
    # 'ERR031544.sort.bam': 'NA'

    _vcf_fields = ('CHROM', 'POS', 'ID', 'REF', 'ALT', 'QUAL', 'FILTER', 'INFO')


    def _format_vcffile(dictreader, vcffile):
'''
Create a pseudo .vcf file based on values read from DictReader instance.
:param dictreader: DictReader instance to read data from
:type dictreader: csv.DictRedaer
:param vcffile: output file.vcf filename
:type vcffile: string
'''
with open(vcffile, mode='w') as writer:
    writer.write('#{}\n'.format('\t'.join(_vcf_fields)))
    output_vcf = []
    for line in dictreader:
        CHROM = line['Chr1']
        # TODO Figure out whether we have zero or one based positioning
        POS = int(line['Pos1'])
        SVEND = int(line['Pos2'])
        INFO = 'PROGRAM=breakdancer;SVTYPE={};SVLEN={}'.format(line['Type'],
                                                               0 - int(line['Size']))
        if line['Type'] not in ['CTX']:
            INFO += ";SVEND={}".format(SVEND)

        # Create record
        output_vcf.append([CHROM, POS, '.', '.', '.', '.', 'PASS', INFO])

    # Sort all results
    output_vcf.sort()
    output = "\n".join(["\t".join(map(str,vcf_row)) for vcf_row in output_vcf])
    # Write record
    writer.write(output)


    if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('-i', '--breakdancertsv', dest='breakdancertsv', type=str,
        help='Breakdancer TSV outputfile')
parser.add_argument('-o', '--outputvcf', dest='outputvcf', type=str,
        help='Output vcf to')

args = parser.parse_args()
main(args.breakdancertsv, args.outputvcf)

复制代码或者github clone到本地文件，执行

    breakdancer2vcf.py -i breakdancer.bam.out -o test.vcf

得到vcf文件后用annovar的区域注释

    convert2annovar.pl -format vcf4 ./breakdancer.out.vcf >test.avinput
    annotate_variation.pl -regionanno -build hg19 -out test -dbtype dgvMerged test.avinput ./