Multiple mapping locations for the same read

[deMan]

If reads can be mapped to several equally best positions, tools like MAQ will randomly choose one position and give the alignment a zero mapping quality and you cannot see all the hits of the read against the reference sequence.

In my case I want to know all the positions of the reads that mapped to those equally best positions.

Does anyone know a tool that can solve this problem?

[Richard Finney]

Easiest way is take reads and run using blat. Blat will report all hits.
If it's just a few reads of interest against a known genome, you can do this on UCSC on-line blat. NCBI's blast will do the job also.

Also, BWA can report alternate hits in XA tag, check out -n and -N parameters to samse and sampe options.

[kmcarr]

If reads can be mapped to several equally best positions, tools like MAQ will randomly choose one position and give the alignment a zero mapping quality and you cannot see all the hits of the read against the reference sequence.

In my case I want to know all the positions of the reads that mapped to those equally best positions.

Does anyone know a tool that can solve this problem?

Bowtie has a number of options which allow you to configure which and how many alignments are reported.

[*#1*]

In some aligners there is a permissivity option that defines to how many locations a read can be mapped to. By default it is usually 1, i.e. once a good match is found, teh read gets mapped to it and then it is sequestrated from the pool of reads and no longer used further. If you set the permissivity parameter to, e.g., 3, the aligner will report up to 3 mapped locations for a read. Just choose your alogner and read documentation to find out what switches to use.