ERCC spike in

Question

[GLFR]

Hello!

I apologise if I am asking a basic question but I was wondering if someone here could clue me in about the role of ERCC spike-in for RNA-Seq?

I've been given a few sets of RNA-Seq data to align to a reference genome and do differential gene expression analysis. I was going to do this via mapping to the reference as opposed to de novo.

I noticed when blasting my over-represented sequences generated from FASTQC that in one sample, I had an over-represented sequence caused by the ERCC spike in. I've tried to understand the role of this in differential gene expression analysis but I'm struggling a bit.

My questions are:

1)Is it normal to present as an over-represented sequence in 1 sample only? 2) Do I need to remove it for mapping and differential expression analysis? 3) If I need to remove it, what's the best way of going about it?

Thank you very much in advance,

Gill

Answer 1

[Devon Ryan]

1. No, that suggests that there were larger problems with the sample showing that (or someone screwed up during library prep...with the former more likely).
2. Nah, at least if you're using human/mouse/"something else common" then the sequences shouldn't align to your genome at any considerable rate.
3. If you wanted to, cat the ERCC sequences to the genome, index that, and align to it.

[GLFR]

Hi Devon,

I'm using A. thaliana, I assume that will very much count as "common".

Thank you so much for your help,

Gill

[Devon Ryan]

I'd think so :)

[zizigolu]

Hi @Devon

I just noticed I have these genes in my raw read counts file

ER-013
ER-028
ER-067
ER-099
ER-109
ER-120
ER-147
ER-154
ER-160
ER-171
ERBB2
ERBB3
ERBB4
ERCC1
ERCC2
ERCC3
ERCC4
ERCC5
ERCC6
ERCC8

I read likely they are ER- series of probes correspond to specific transcripts within the ERCC RNA spike-in. I have also noticed one of them is among my differentially expressed genes, so should I remove them before any quantitative procedure?

Thank you

[Devon Ryan]

I'm not familiar with the ER- entries.