unexpected cDNA product during library prep

[amango]

Attached are Bioanalyzer traces from various points in my illumina library preparation protocol for RNA-seq, modified to sequence only 3’ transcript ends. Graphs are from different samples since limits in starting amounts don’t allow me to validate any singe sample at each of its steps.

My question is: what are the high molecular weight molecules between 1000-4000bp in my cDNA samples (graphs D1 & D2)? I would think that if the mRNA template is fragmented, the lack of 1-4k length rna templates (avg fragment length <200bp, graph C) would preclude synthesis of long fragments. And RNase H treatment in my protocol should eliminate long RNA contamination. It looks like there are some cDNAs of appropriate length (i.e. <200bp). Thanks for any ideas!

The simplified protocol is:
1. From total RNA (graph A, below), isolate mRNA (graph B) using oligo d(t) dynabeads
2. Fragment mRNA using Ambion RNA fragmentation reagents (graph C)
3. Select poly-A fragments from fragmented mRNA using oligo d(t) dynabeads
4. Synthesize first strand cDNA using oligo d(t) primers and Invitrogen Superscript II
5. Synthesize second strand cDNA using DNA pol I and RNase H (graphs D1 & D2)
6. Blunt-ending, adenylation, adapter-ligation, size-selection, pcr enrichment, etc

[Nik]

Hi,

I've seen this as well. I'm trying to work out what this is but can't help you at the moment but it might be nice to know you are not alone. will let you know if some test work out and give me an idea/answer.

[greigite]

Hi,
Reviving this old thread to see if any explanation was found for the long cDNA molecules. We are seeing exactly the same thing with a similar, but slightly different protocol (mRNA isolation, RNA amplification, fragmentation, then Invitrogen super-script ds cDNA synthesis).

[Yke]

Hi,

I wonder if anybody made progress in this issue. It seem to me that at least 4 different people had the same problem. I am having exactly the same problem and would like to solve it. Any clues? Did anybody run an agarose gel from the cDNA instead of the bioanalyzer? It could be a detection problem...
Thank you!

[greigite]

linear amplification of RNA is generating long molecules

Taking a closer look at our data it seems that the long molecules are generated in the RNA linear amplification step (using Ambion's Message-Amp II-Bacteria kit), not the cDNA synthesis. The cDNA synthesis actually tends to decrease the proportion of molecules in the 200 bp-1kb range relative to the RNA though the proportion between 100 and 200 bp is the same.

The problem with the aRNA is substantial though: only 7% of molecules were above 1 kb in the original mRNA and 47% were above 1 kb in the amplified sample. This smells like concatemerization to me but I don't know. We are cloning some long cDNA to check.