Novel Transcript discovery in paired samples

[astar]

Hi all,

I hope you can advise me on a tool for RNA-seq data analysis for paired samples.

I will analyse data from RNA-seq experiment. We have 2 samples for each of our 3 patients (before and after treatment).

We would like to investigate two things:
1) change in expression of known transcripts between before and after treatment sample

2) identify novel transcripts and compare their expression between before and after treatment sample

We plan to align reads with the use of TopHat and then we considered using DESeq2 as it allows to take under consideration the paired design.
What workflow would you recommend for such a design?

I will appreciate any advice.

Best,
astar.

[ml439]

Hi Astar,

I'm not an expert in the field, but I have just done something similar with my data so here's my answer.

For your point 1) I will map with tophat2, count your reads using HTSeq, providing HTSeq a GTF with known transcripts (you can get that from Ensembl or UCSC) and then use DESeq2 with those reads.

For your point 2) I have just used a tool called StringTie, which has performed very well in my hands to annotate new transcripts. StringTie provides a GTF file, that you can use afterwards for counting with HTSeq and DESeq2.

Have to mention that I've only done point 2, but that has identified known transcripts as well as new ones all at once. If you want to separate the analysis to report old and new transcripts independently, you'll have to filter the output at the end.

I hope this is helpful and not too late for you,

good luck

Miriam

[astar]

Hi Miriam,

thank you very much for your reply and your advice.
I will try it out for sure

Best,
astar