bam-rewrap.1

.\" Process this file with
.\" groff -man -Tascii bam-rmdup.1
.\"
.TH BAM-REWRAP 1 "SEPTEMBER 2013" Applications "User Manuals"
.SH NAME
bam-rewrap \- rewrap alignments in BAM file to real target length
.SH SYNOPSIS
.BI "bam-rewrap [" chrom : length " ...]"

.SH DESCRIPTION
.B bam-rewrap
reads a BAM file from standard input and writes a BAM file to standard
output where every alignment to a subset of target sequences has been
wrapped to the targets actual length, obtained from the command line.

The idea is that a circular reference sequence has been extended
artificially to facilitate alignment.  Now the declared length in the
header is wrong, and some alignments overhang the end. 
.B bam-rewrap
splits
those alignments into two, one for the beginning, one for the end of
the sequence, then soft-masks out the inappropriate parts.  Alignments
falling completely behind the actual end of the target sequence are
wrapped to natural coordinates.

.B bam-rewrap
tries to fix the map quality (MAPQ) for the affected reads as follows:  if
a read has zero map quality, meaning multiple equally good hits, 
.B bam-rewrap
checks the 
.I XA
field.  If it reports exactly one additional alignment,
and it matches the primary alignment when transformed to natural
coordinates, 
.I XA 
is removed and 
.I MAPQ
set to 37, indicating a unique hit.  This logic is not standardized and
is only known to work if alignments were produced by
.BR bwa .
It may or may not make sense for other aligners.
 
.SH OPTIONS
.IP "chrom:length"
Indicates that the length of target sequence
.IR chrom " is " length .
This option can be repeated and each will cause alignments to the
specified target to be wrapped.  A unique prefix of the sequence name is
sufficient.

.SH AUTHOR
Udo Stenzel <udo_stenzel@eva.mpg.de>

.SH "SEE ALSO"
.BR biohazard (7)