ComputeReadSpanCoverage

Geraldine_VdAuwera

1. Introduction

The ComputeReadSpanCoverage walker traverses a set of BAM files to generate genome-wide statistics.

The read span coverage is the count of bases in between two paired-end reads, not counting the lengths of the reads themselves. For fixed-length reads of length L with ungapped alignments, this would be InsertSize - 2*L. The read span coverage is used as an estimate of the power for detecting breakpoints using read pairs. This estimate assumes a model where the aligner is unlikely to align a read to a breakpoint unless the breakpoint is close to the end of the read.

Read pairs where the ends align to different sequences are never counted.

Read span coverage is computed and reported for each readgroup, but the output is keyed by sample and library to allow easy roll up.

See also MergeReadSpanCoverage.

2. Inputs / Arguments

• -I <bam-file> : The set of input BAM files.

• -md <directory> : The metadata directory. Insert size histograms are loaded from the default isd.hist.bin file in this directory. This argument is also used to load a default list of excluded read groups.

• -maxInsertSize <n> : Read pairs with an insert size greater than n are not counted in span coverage.

• -maxInsertSizeStandardDeviations <sd> : Read pairs with an insert size greater than the median plus sd robust standard deviations are not counted in span coverage.

3. Outputs

• -O <span-coverage-file> : Tab delimited output file (default is stdout).