softwares - splicing

[cedance]

Hi,

I am working on RNASeq data on paired reads. I use tophat for splice junction mapping. I would like to know other tools developed, to compare and contrast the results from both softwares...

Thank you.

[kopi-o]

Here are some I've tried
MapSplice
SpliceMap
RUM

[cedance]

Thank you very much. I'll checkout all of them! However, in your opinion, which one would you recommend?

[kopi-o]

It's hard to recommend one over the others without extensive testing, which I haven't done. In my particular case, RUM seemed to recover the most spliced alignments, but that could vary depending on your experiment, program parameters, etc etc...

[edge]

Hi kopi-o, how do you identify that RUM seemed to recover most spliced alignments?
Got any specify rules to identify it?
Many thanks for advice.

[kopi-o]

I look for strings with containing "N" in the CIGAR string in the resulting alignment SAM file.

[edge]

thanks a lot, kopi-o.
Do you ever try Tophat for mapping purpose?
Do you use the RUM output result for assembly?
As I know, RUM just focus on optimum mapping result but it don't provide assembly
I will prefer to undergo mapping then assembly the RNA-seq data.
thanks for any advice ya.

[kopi-o]

Hi,
Sure, I use mostly TopHat, but in this case I ran RUM (and some other packages like MapSplice, SpliceMap) for comparison. If you want to do reference-guided assembly, you can try Cufflinks or Scripture.

[tboothby]

Anyone know any programs for this that don't require a reference to map against?

[kopi-o]

E.g. Trinity, Velvet combined with Oases, and TransAbyss.

[edge]

Hi kopi-o,
I got run through RUM with my RNA-seq data set.
Unfortunately, I can't really get the meaning of each output result
After I count back the total read in RUM.sam, it seems like all the raw read are alignment to either genome or transcriptome.
Thus I not really sure what is the proper usage of RUM output result.
Do you mind to share more about which file or info that you used for assembly after then?
Actually my main purpose just wanna used RUM to improve the mapping result and then use those raw read for de-novo assembly of my RNA-seq data.
Many thanks for advice.

[sphil]

Hey,

mgene.ngs can also do de novo gene finding. It uses SVM and HMM for that.

best

philip


(available at:http://www.fml.tuebingen.mpg.de/raetsch/suppl/mgene )

edit: it's for DNA sequences... sorry

[tboothby]

E.g. Trinity, Velvet combined with Oases, and TransAbyss.

Yea, so I assembled sequences using Trinity which gives me an output file with all my different transcript isoforms.

My question is, does anyone know an efficient way to sort these isoforms? I am particularly interested in finding alternatively spliced transcripts.