recover cDNA from total RNA for RNAseq?

[mapplebe]

We have a large set of drosophila samples that we are in the process of generating transcription profiles of using illumina RNAseq. Originally, a large number of these samples were sent to a private contractor, but alas, their initial results were not so good so we demanded them back and we have been slowly doing them in-house ever since.

Additionally, we were originally going to run these on microarrays instead of using illumina RNAseq. Because of this, the private contractor returned about 25 of these samples as ds-cDNA generated directly from total RNA, which is the protocol for microarrays, but now my boss desperately wants me to figure out how to separate the mRNA-cDNAs so we can run them on illumina. Anyone have a clue how to do it?

The only idea I had was to generate "1st-strand" mRNA/cDNAs using poly-T primer and a DNA polymerase, running amplification for 2-3 cycles. Then if I can get rid of the poly-T primer without getting rid of the generated ss-cDNA (PEG precipitation?) I could use poly(T) beads to enrich for the ss-cDNAs. Thoughts?

[kerplunk412]

I think it depends on how the ds-cDNAs were generated. If they were generated with random primers, many of your mRNA/cDNAs may not have A tails, so the strategy you describe may not work. If they were generated using oligo d(T), which I think is more standard, the strategy you described should work, IMO.

[pmiguel]

If you have ds cDNA, why not just use an Illumina DNA prep kit?

--
Phillip

[kerplunk412]

Yeah, good point. I think his concern was that he didn't want to sequence a bunch of rRNA reads, but if the cDNA was created with oligodT priming that shouldn't be much of an issue.