API

class dysgu.DysguSV(ref_genome, bam, sample_name=’sample’, **kwargs)

This class is the main interface for calling structural variants using dysgu. To initialize DysguSV, provide a reference genome and bam file, via the pysam library. It is also recommended to provide a sample_name which will be used when saving results to a vcf file.

import pysam
from dysgu import DysguSV

# open a reference genome and alignment file using pysam
bam = pysam.AlignmentFile('test.bam', 'rb')
genome = pysam.FastaFile('ref.fasta')

# Initialise dysgu
dysgu = DysguSV(genome, bam)

# Call SVs at a genomic location
df = dysgu("chr1:1-1000000")

To options can be provided during initialisation as key-word arguments:

dysgu = DysguSV(genome, bam, min_support=5, mq=20)

__call__(region, sort_df=True)

Call SVs using dysgu

dysgu = DysguSV(ref, bam)
df = dysgu("chr1:10000-50000")

# Using a pysam iterator
df = dysgu(bam.fetch("chr1", 0, 500000))

apply_model(df)

Apply a machine leaning model to the dataframe. The model configuration is determined by the options set on the DysguSV class. For example, to use a non diploid model, first set diploid=False:

dysgu = DysguSV(ref, bam)
dysgu.set_option("diploid", False)
df = dysgu("chr1:10000-50000")

call_bed_regions(regions)

Call SVs from a list of bed regions. Note, bed regions should be sorted by genome starting position, and be non-overlapping. To create a suitable input for this function see dysgu.load_bed() and dysgu.merge_intervals() functions.

from dysgu import load_bed, merge_intervals

# load, merge and sort intervals
bed = load_bed('test.bed')
bed = merge_intervals(bed, srt=True)

dysgu = DysguSV(ref, bam)
df = dysgu.call_bed_regions(bed)

set_option(option, value=None)

Change option(s) for dysgu.

dysgu.set_option("min_support", 10)

# Or provide a mapping of arguments:
dysgu.set_option({"min_support": 10, "mq": 20, "min_size": 100})

to_vcf(dataframe, output_file)

Save dysgu SV calls to a vcf file

with open(path, "w") as out:
    dysgu.to_vcf(passed, out)

dysgu.dysgu_default_args()

Returns the default arguments used by dysgu

dysgu.load_dysgu_vcf(path, drop_na_columns=True)

Load a vcf file from dysgu

dysgu.merge_dysgu_df(*dataframes, merge_distance=500, pick_best=True, add_partners=True)

Merge calls from dysgu. Input is one or more dataframes with dysgu calls.

dysgu.merge_intervals()

Merge a list of intervals, the expected format is a 3-tuple e.g. (chromosome, start, end). If add_indexes is set to True, merge_intervals expects a 4-tuple with the last item corresponding to an index variable

>>> merge_intervals( [('chr1', 1, 4), ('chr1', 2, 5), ('chr2', 3, 5)] )
>>> [['chr1', 1, 5], ['chr2', 3, 5]]

>>> a = [("chr1", 1, 10, 0), ("chr1", 9, 11, 1), ("chr1", 20, 30, 2)]
>>> merge_intervals(a, add_indexes=True)
>>> [('chr1', 1, 11, [0, 1]), ['chr1', 20, 30, [2]]]