rRNA and adapter removal bondaries in RNA-seq HiSeq 3000 Illumina

[marnal]

Hello

I am new in RNA-seq analysis and it is difficult for me to find out info about quality assessment in the matter. I did an analysis with fastqc and MultiQC in a Paired End, 125nts (2x125), HiSeq3000 experiment. I obtained these two plots (attached) and I am not sure if a rRNA or adapter removal is required prior to alignment.

In the case of the adapter content plot it looks like all samples have adapters included, in fact none of the samples passed the QA. In the case of the plot with %GC content (I read it can be used to detect a possible rRNA enrichment) I see a proportion of reads having a GC content higher than 60%, but I am not sure if it is enough to confirm a high presence of rRNA that could affect to posterior analysis.

Any help will be useful!

Thanks

Magda

[GenoMax]

You should scan/trim your sequences with a suitable program (bbduk.sh from BBMap, trimmomatic, cutadapt etc.). While some rely on aligners to soft-clip adapter sequences (and they likely will), trimming them beforehand ensures that your initial data is clean.

Presence of rRNA is a separate issue. If you feel that your data has rRNA reads in it then the best solution is to align a sample (or whole data) to rDNA repeat for your organism. While rRNA reads would not directly interfere with counts if you have uneven levels of them across samples then they can indirectly affect your results.

[marnal]

Thanks for your quick answer, it was very useful! Just a last question, could you recommend me a tool or database I could use to make the alignment with the rDNA repeats of an organism?

Thanks

[GenoMax]

What organism are you working with?

[marnal]

We are working with human

[GenoMax]

You can get the human rDNA repeat from here. If you do see contamination you can use bbsplit.sh from BBMap suite to separate those reads from rest of your data, if you wish.