[miso-users] Required SAM format

Yarden Katz yarden at MIT.EDU
Fri Aug 16 14:29:00 EDT 2013


Hi Alex,

See comments below:

On Aug 13, 2013, at 6:56 AM, Alexander Kanitz wrote:

> Hi everyone,
> 
> I am planning to use MISO but I want to be flexible with the aligner I use (read: don't want to use Bowtie/TopHat). 
> 
> Since I did not find any detailed specs regarding this issue, I have a few questions:
> What does MISO need out of a SAM/BAM file? Specifically, which tags are considered and which are required (assuming all others will be ignored)? 

MISO tries to not rely on too many binary SAM/BAM flags since as you imply many different aligners follow slightly different conventions.  

The main flags used as strandedness if you have stranded data and paired-ness/orientation of reads if you have paired-end data.  There's a flag to MISO that can tell it to group read pairs by ID, e.g. assuming that the reads have the format:

readX/1
readX/2 

or something similar that makes it clear that they are both mates of the same sequenced molecule. 

> And most importantly: in what format (one line, separate lines, CIGAR string, MD tag, etc) do split/spliced alignments have to be represented?

CIGAR string.  I can add a note to the manual clarifying this. 


> 
> I would be very grateful for some help here and ideally a few lines of a MISO-compatible SAM file including some split/spliced alignments if possible (or a link to such a file).

Any BAM file produced with Tophat and Bowtie 1 should be an example of a MISO-accepted BAM, and we normally use Tophat around here though have users that use other aligners.  I am traveling and don't have access handy now, but it should be easy to find. 

One thing to note: MISO ignores insertions and deletions as they do not match the annotation -- so any reads with insertion/deletions in their CIGAR strings will be discarded.

> 
> Thanks a lot!
> 
> Best,
> Alex
> 
> 
> -- 
> Alexander Kanitz, Ph.D.
> 
> Winterthurerstrasse 358
> 8057 Zurich
> Switzerland
> 
> Phone: +41 76 278 39 36
> Email: alexander.kanitz at alumni.ethz.ch
> _______________________________________________
> miso-users mailing list
> miso-users at mit.edu
> http://mailman.mit.edu/mailman/listinfo/miso-users




More information about the miso-users mailing list