Too many reads mapping towards intronic region

[sanush]

Hi all
We just finished our first RNA seq experiment in SOLiD platform and used partek to map our reads. We see that only 30% maps to exonic and remaining of our reads mapping to intronic and intergenic regions.

We are trying to figure out the reason for high numbers mapping to intronic regions. Any advice or comment would be appreciated.

Thanks

[NextGenSeq]

How did you isolate the RNA? Was it mRNA or total RNA? If total RNA did you remove 18S and 28S rRNA? Did you DNAse treat your RNA to digest contaminating DNA away?

If the reads mapping to intronic and intergenic regions are high quality sequence, I suspect contaminating genomic DNA.

[pmiguel]

Hi all
We just finished our first RNA seq experiment in SOLiD platform and used partek to map our reads. We see that only 30% maps to exonic and remaining of our reads mapping to intronic and intergenic regions.

We are trying to figure out the reason for high numbers mapping to intronic regions. Any advice or comment would be appreciated.

Thanks

What organism?
Was this done on Ribominus rRNA depleted samples? If so, then your results are as expected.

--
Phillip

[sanush]

Hi
Thanks for the reply. These are human cell line samples total RNA sample. We did double ribominus treatment to make sure 18s and 28s are removed. We checked it on bioanalyzer pico kit, but we did not do the DNase treatment. Even we suspected it could be genomic DNA contamination but is there any other possibility.

Thanks

[Chipper]

Most likely it is a combination of pre-mRNAs, rRNA (you can't expect all to be removed, especially if you have some degradation of the sample), snRNAs, lincRNAs, unannotated utrs etc. I am not familiar with Partek, but the number of mismatches you allow in alignment will also affect the % in exons, low-quality reads are less likely to be correctly placed. We had around 50 % in refGene exons using bwa.

[pmiguel]

Hi
Thanks for the reply. These are human cell line samples total RNA sample. We did double ribominus treatment to make sure 18s and 28s are removed. We checked it on bioanalyzer pico kit, but we did not do the DNase treatment. Even we suspected it could be genomic DNA contamination but is there any other possibility.

Thanks

Okay, then I think your results are reasonable. A large percentage of the genome is transcribed. We normally do not see it because we either prime from the polyA tail or do polyA+ isolations.

I don't think DNase is a big issue. In fact, unless you are sure you are doing a complete DNase digestion, you are probably better off not doing it. The overhung adaptors used in the SWTAK would heavily favor single stranded RNA/DNA. Since the majority of DNA will be (1) huge, (2) not fragmented by RNAseIII and (3)double stranded, I do not thing much DNA would get cloned. If you hit your RNA with DNase, you would likely bring some it down into the size range where it would be clonable--if it had non-blunt ends.

--
Phillip

[pmiguel]

Hi
Thanks for the reply. These are human cell line samples total RNA sample. We did double ribominus treatment to make sure 18s and 28s are removed. We checked it on bioanalyzer pico kit, but we did not do the DNase treatment. Even we suspected it could be genomic DNA contamination but is there any other possibility.

Thanks

I got your PM, but would prefer to keep the discussion here. You are asking why I think that 2/3rds of transcripts would map outside coding regions. I guess you could do a literature search on non-coding RNAs if you want the full story. Here is a good place to start: http://www.sciencemag.org/cgi/conten.../308/5725/1149

But a high percentage of the human genome is transcribed. You have all the small RNAs, pre-adenylated non-mature mRNAs with introns not yet spliced out. The intronic RNA post-splicing. Then you have transposable elements--many lurking in introns. Also some regulatory regions (such as enhancers) are transcribed. And there is probably other stuff transcribed for reasons we do not yet understand.

If you are not interested in these transcripts, use polyA+ RNA rather than ribominus.

--
Phillip