reference.py 文件源码

def load_primary_contigs(reference_path):
    '''Load set of primary contigs for variant and SV calling from reference_path.
       If now primary_contigs.txt file is specified, return all contigs. If  reference_path
       is a known 10x reference genome and has no primary_contigs.txt, filter the known bad contigs '''

    if not reference_path is None and os.path.exists(get_primary_contigs(reference_path)):
        # If we have a primary_contigs.txt file, use it
        with open(get_primary_contigs(reference_path), 'r') as f:
            primary_contigs = set([line.strip() for line in f.readlines()])

    else:
        # Default is to include all contigs
        # Otherwise implement the old contig filters
        ref = open_reference(reference_path)
        primary_contigs = set(ref.keys())

        if is_tenx(reference_path):
            primary_contigs = set(chrom for chrom in primary_contigs if not ('random' in chrom or 'U' in chrom or 'hap' in chrom or chrom == 'hs37d5'))

    return primary_contigs