reads that align to multiple places

[Bogdan Tanasa]

Dear all,

please could you suggest an appropriate combination of parameters I can use in order to output :

<> the reads that align uniquely
<> the reads that align to less than 10, or less than 100, 1000 places
<> if a read aligns to multiple places, one random place can be chosen

thanks a lot,

Bogdan

[Nathan Clement]

See below.

Nathan

On Nov 4, 2011, at 3:00 PM, Bogdan Tanasa wrote:

> Dear all,
>
> please could you suggest an appropriate combination of parameters I can use in order to output :
>
> <> the reads that align uniquely

There's a flag that's included, but not extensively tested, and therefore not listed in the parameters list. If you include "-u" as a parameter, it will not report the sequence as being aligned if it matches a sequence that occurs in more than one location. In other words, if a genomic sequence aligns to the read with a score higher than the threshold score (say, 90%), but occurs at more than one location in the genome, the read will not align.

> <> the reads that align to less than 10, or less than 100, 1000 places

Use the "-T" flag to change the number of matches (default is 1000).

> <> if a read aligns to multiple places, one random place can be chosen

This currently has not been implemented in GNUMAP, but would be relatively easy to do so, even as a post-processing step.

>
> thanks a lot,
>
> Bogdan