3'UTR- Length

[tr7]

Dear all,
I am very new to RNAseq, thus I am sorry if this is a trivial question. I need information about the length of 3'UTRs. I have run Tophat for anlignment and Cufflinks to get transcript expression values, but I only got exon information. How can I get information about the UTRs?
I would be very glad about any help!
Thanks a lot!

[308350576]

some problem of 3P-seq,anyone come to discuss!

Is there anyone doing 3P-seq,one protocol of 3' UTR region.I have some problem of the reagent of the protocol,anyone can discuss it?

[syfo]

Dear all,
I am very new to RNAseq, thus I am sorry if this is a trivial question. I need information about the length of 3'UTRs. I have run Tophat for anlignment and Cufflinks to get transcript expression values, but I only got exon information. How can I get information about the UTRs?
I would be very glad about any help!
Thanks a lot!

Unless your RNA-seq protocol was targeting non coding or small RNAs, you are mostly sequencing mRNAs that have been spliced (typically the case if rRNAs have been filtered out using polyA selection), so your reads come from exons. UTRs are the noncoding parts of the exons. They are located at the extremities of the mRNAs: upstream of the start codon for the 5'UTR, downstream of the stop codon for the 3'UTRs. As you see, it all depends on the coding part of the mRNA (the CDS), which most of the time requires the knowledge of the complete exon/intron structure of the RNA to be identified.
So I would say that any investigation about UTRs requires the preliminary identification of full-length mRNAs, which is not necessarily achieved by using TopHat+Cufflinks on traditional RNA-seq data. You might want to take a closer look at topics like "genome annotation", "gene structure prediction" or "gene finding" to get a better idea.

This said... a quick and dirty hack to start investigate your data anyway would be to extract the potential transfrags you got from Cufflinks (after joining the ones that are predicted to flank introns) and run some ORF finder on the sequences that are long enough. You might identify a couple of STOP codons, and therefore 3'UTRs.. but this would be extremely incomplete anyway.
Another strategy is to perform a "de novo" assembly of your RNA-seq data to get mRNA sequences and look for ORFs/UTRs in those.

I am assuming your genome is not annotated of course, otherwise you already know where 3'UTRs are -in which case you might just be interested in computing their coverage.

Good luck,
s.