Differential expression, pair end

[kenietz]

Hi all,
as i am relatively new to the field i have a couple of questions:

1. Is it possible to do RNA-seq analysis when starting from cDNA? The lab guys dont have enough RNA so they wanna amplify first.
In my opinion that should not be a problem but i am not sure. On the web i found only one post where someone stated they started from cDNA.

2. Is it possible to do a Differential expression(DE) with Illumina PE data? For 'denovo' which i should also do is perfect but i am not sure about the DE.

Thank you for your time and any help is appreciated.

[ScottC]

Yes, you can start from cDNA. How much cDNA do you have?

[kenietz]

I do not know how much cDNA they have. But i suppose they will have enough after amplification. Am i to assume that i would be able to do and DE with Illumina PE data from cDNA?

[pbluescript]

I've done differential expression analysis using paired end sequencing after starting with hundreds of picograms of RNA after amplifying it and converting it to cDNA. It worked well, even though I had a lower than usual number of reads that mapped.

[kenietz]

I see, thank you very much for the information. I had the feeling it should work but was not sure. Btw, which tool did you use for DE analysis? I saw there are many out there like DESeq, edgeR, NOISeq, FET etc?

[pbluescript]

My most commonly used tool is cuffdiff since it is very easy to include multireads. However, I've found the best results when I use more than one tool and compare the results. I'm a bit partial to rsem and baySeq, but I've found DESeq and edgeR to be useful too.

[kenietz]

thanks. i was reading a paper "Differential expression in RNA-seq: A matter of depth" by Sonia Tarazona and there i saw that NOISeq and FET seem to perform the best. But cuffdiff is most widely used i think. So i suppose i will try cuffdiff and NOISeq and see whats the result.