Do I need Illumina Demultiplexing for RNASeq?

Question

[Hashirama]

Dear Community,

I am a little bit confused analyzing my RNASeq data. However, I sent 12 samples to the sequencing facility, 6 control samples and 6 knockdown samples. However I only got 3 files back from the sequencing facilty:

lane1_R1.fastq.gz 
lane1_R2.fastq.gz 
lane1_R3.fastq.gz 

Why do I get 3 files back, when I sent 12 samples to the facility? Thats really confusing... Do I maybe need to demultiplex my data? I tried to analyze the data without demultiplexing with the following workflow on GALAXY so far:

FATSQC -> Trimmomatic -> HISAT2 -> StringTie -> SeSeq2 -> ...

However StringTie to read the quantification does not work really.. How can I solve this?

The descrption of RNASeq data I got from the seqcuencing facility says:

"1,2,3,4,5,6,7,8,9,10,11,12 single + index nn RNAseq Y50N,Y8,Y8 1 1 Galaxy no_demux "

However, I am really confused and would be glad for any help from you guys!

Thanks, Hashirama

Answer 1

[lieven.sterck]

without knowing what exactly happened , you should have (start with) 12 files, one per sample you sent and that three times: 36 in total thus (12 samples x3 reps)

So, yes, my guess is also that you will need to demultiplex them. There should be tools in galaxy as well to achieve that.

Analyzing the three files you have as such will certainly not work

[Hashirama]

Thanks lieven for your quick and helpful answer. It's really confusing that I only got 3 fastq files... So I tried demultiplexing on Galaxy - however, I need to add some "Illumina index read" files for the analyses. I found an excel datafile with "i7 and i5 Index Sequences" one the website of the sequencing facility but I don't know how to use them for the further analysis. Does anyone have an idea?

[Ventrilocus]

Hi there, Hashirama. Very often, sequencing facilities enable the Real-​Time Analysis (RTA) Software that can de-multiplex straight out of the machine. This was not your case and thus, you need to demultiplex yourself. I have no idea whether you can do it with Galaxy. Sabre (https://github.com/najoshi/sabre) is a popular choice. Make sure you contact the facility to help you with the demultiplexing (they do this on a daily basis). If the so called facility is actually a commercial enterprise, go to your receipt and check whether demultiplexing was part of the inquiry (they may charge you extra for demultiplexing).

[lieven.sterck]

indeed, you will need info from your sequencing facility to get to know the barcodes they used. (perhaps it is somewhere in the info you got from them?)

Answer 2

[GenoMax]

This is irresponsible behavior from sequencing facility if they did not demultiplex the data before giving it to you. You should go back to them and ask them for demultiplexed data in format that lieven.sterck described below. This is something trivial for them to do. Did you make the libraries yourself or did the facility make the libraries? In either case it is simple for them to do the demultiplexing than have you try to work this out on your own.

[Hashirama]

No the facility did the libraries but they refuse to do the demultiplexing saying that I have to to it on my own. I thinks that's quite irresponsible like you said! However, what do I need to do the demultiplexing on my own. The libraries?

But anyway thanks for your answer. Now I know that its not my fault that I can't finish the data analysis :-)

[lieven.sterck]

the minimum you will need is the list of barcodes used in the multiplexing (each barcode should correspond with one of your sample)

I can only confirm that is very irresponsible behavior of that facility, as GenoMax said, it's a straightforward analysis for them.

[GenoMax]

facility did the libraries but they refuse to do the demultiplexing

I hope they gave you a list of index --> SampleID combinations otherwise you are not going to be able to figure this out (well technically you can but you will not know which file represents what sample).

You can see which indexes are present in your sample by using the code here: Demultiplexing reads with index present in the labels