RNA-seq pipeline

[himanshu04]

Hey everyone,
I am new to this field and I am really interested in learning more about it. I have a project to do in which, I have RNA-seq data and I have to do the following:
1.how do you think you would try to go about figuring out how to identify short non-coding versus coding transcripts?- I think this could be done by doing a Blast x using Blast2go.
2.How would you try to identify translocations/alternative splice variants? - I am not very sure about this one but should I be finding out the differential expression or protein isoforms for this using Tophat and Cufflinks.??
3.Then to validate your results, to obtain statistical significance? - I think I would have to use Bioconductor packages DEseq/edgeR.
Also, can anyone give a better suggestion or some guidelines/steps for the same. Any help is much appreciated.
Thanks.

[himanshu04]

Anyone can help me with this please?

[turnersd]

As for #3, using a different tool/algorithm to analyze your data isn't independent validation. There's a long discussion in other threads here on the relative merits of DESeq vs something like cufflinks. Some of the older posts are out of date - coming from a time before cufflinks allowed biological replicates.

[himanshu04]

Thanks a lot for your reply. I did take a look at those posts but since I am new and trying to do it for the first time just wanted to know whether this approach is correct?.
Thanks.

[turnersd]

To get at splicing, using cufflinks to look at different isoform expression or splicing load is a good place to start. But you can't then look at results from running DESeq on raw counts as independent validation. You need new data for that, or functional studies, qPCR, etc.