Reads mapped to exonic, intronic and intergenic regions

Question

After the alignment step I checked the rnaseq metrics of all the samples. Among 40 samples three samples show high percentage of reads mapped to intronic regions. What could be the reason?

Samples   Exonic            Intronic    Intergenic
Sample1 545479 (12.8%)  2512309 (58.8%) 1217201 (28.5%)
Sample2 372836 (8.3%)   2556934 (56.8%) 1573032 (34%)
Sample3 529618 (7.3%)   3934615 (54.5%) 2750720 (38.1%)

I also see that 35-40k genes among 58k genes having "zero" read counts. Is this due to contamination? What could be the reason? What does reads mapping to exonic, intronic, intergenic tell?

Update: I used hisat2 for alignment. I used human genome "grch38_snp_tran" from hisat2 website. Libraries were generated using ribosome depletion kit.

How did the overall alignment rate of those samples compare to the others that had more reasonable intronic alignment rates? — Devon Ryan, May 07 '18 at 19:40
I see that all the 37 samples have overall alignment rate between 90-96% and for those 3 samples it is 69%, 79%, 70% overall alignment rate. — stack_learner, May 07 '18 at 20:12
U mean those 3 samples are degraded? What I should do now? Should I remove them for further analysis? — stack_learner, May 07 '18 at 20:34
It’s a likely cause, but regardless you should exclude them. — Devon Ryan, May 07 '18 at 20:35
Sure. the high percentage of mapping to intronic region is due to degradation? — stack_learner, May 07 '18 at 20:55
Could you add the quality metrics of these files and other of the same run to the question? — llrs, May 08 '18 at 06:58
There is the potential that these samples had low quality RNA or genomic DNA contamination (depending on the sample type and RNA isolation method) — GWW, Jan 09 '19 at 15:15

score 3 · Answer 1 · answered Jan 09 '19 at 15:06

3

As @DevonRyan mentioned, it's very likely that those samples were degraded, which is good justification for excluding them from subsequent analysis.

answered Jan 09 '19 at 15:06

Daniel Standage

5,080
15
50

score 2 · Answer 2 · answered Jan 10 '19 at 12:26

I would expect at least 30% of reads from a total-cell, ribo-depleted RNA-seq to be exonic. Less suggests something when wrong.

As well as degradation, another explanation would be contamination with genomic DNA.

RNA obtained from the nuclear or cytoplasmic fraction of the cell might have a exonic content lower than 30%, while poly-A selected RNA would be expected to have higher, more like 66%.

Reads mapped to exonic, intronic and intergenic regions

2 Answers2

Linked