Has anyone tried RUM for aligning/counting Illumina RNA-Seq data?

[fabrice]

2



The publication for this has just arrived: http://rna-seqblog.com/data-analysis...ed-mapper-rum/

Based on speed and accuracy, it looks like a contender. Comparable to GSNAP in the accuracy stakes whilst being considerably faster.

I'm, wondering if anyone has trialled it yet? I ran a simulated data set through it just to see how it ran and it went both fast and smoothly. I am unsure as to how the read counts are calculated for closely related isoforms (i.e. how it distinguishes between them). I contacted the author about it but got no response.

Has anyone else looked into this? Can you offer any thoughts?

Thanks in advance.

[twu]

Hi Fabrice,

I am the developer of GSNAP. Regarding the speed of GSNAP, the latest version (starting with 2011-08-15) is about 4-8 times faster than the one that was tested in that Bioinformatics paper. The latest version of GSNAP uses k-mer sizes of 15 by default, which accounts for the speed up over the previous k-mer size of 12.

In addition, the latest version is more accurate, at least for paired-end reads, since I have incorporated the GMAP algorithm to handle difficult cases involving multiple indels or splicing.

As always, you can download the latest source code for free at http://research-pub.gene.com/gmap.

Regards,

Thomas Wu

[cjp]

Hi Thomas,

I'm trying to look at both RUM and gsnap for RNA-seq alignments and compare to TopHat. gsnap looks like really nice software and I noticed user lh3 recommended it! What command line options would you recommend to align/map RNA-seq to human/mouse genomes for 1) Illumina 36bp single-end reads and 2) Illumina 76bp paired-end reads.

Thanks, Chris

[twu]

Hi Chris,

I tend to run GSNAP myself with mostly default options. However, for RNA-Seq, you need to tell the program to find novel splices with "-N 1", otherwise the program will think you have DNA-Seq data. You will also get better results at the ends of reads if you prepare a known splicesites file (see the README file for details), and use that with the "-s" flag to make use of known splice sites.

You also may want to provide some reasonable value for an allowable number of mismatches with the "-m" flag. If you don't provide it, then GSNAP will pick a value that allows it to compute quickly, but it might pick a value too low for 36-mers. You might therefore select "-m 3" or "-m 4" or so for the 36-mers. For 76-mers, GSNAP will explore up to "-m 4", which is probably fine, or you could specify "-m 5".

Regards,

Tom

[cjp]

Thanks for that Tom,

A bit later, I also found some other links where other users that had different command line options:

http://openwetware.org/wiki/Wikiomics%3aRNA-Seq#GMAP.2FGSNAP

http://seqanswers.com/forums/showthread.php?t=14368

Chris