Read length after adaptor trimming

[ea11]

Hi all,

I have received some RNA-seq data and after looking through the data, it seems that the adaptors have not been fully trimmed.

I used grep with a portion of the adaptor sequence one my fastq files and the TruSeq Illumina universal adaptor seems to be present in my reads as well as the TruSeq adaptor Index sequences.

After trimming with cutadapt, I ran the trimmed file through fastqc again and it seems the sequence length distribution is no longer the same for all the reads.
Prior to trimming, all the reads were 101bp, and now there are reads that are shorter and I have attached the graph I get for the trimmed reads.

Previous RNA-seq data I have received have been trimmed already, so I am new to trimming adaptor sequences. Below is the code I used for trimming the adaptors.

cutadapt -a GATCGGAAGAGCA -g GCTCTTCCGATCT -o "PATH to trimmed file" "PATH to file"

Is it normal to have non-uniform reads after trimming or do they need to all be uniform length for downstream alignment and analysis?

Thanks

[blancha]

Yes, it's normal.

The length of the fragments selected during library preparation will vary, and won't always be exactly the same.
The size of the inserts between the adapters will therefore vary.
For the shorter inserts, if the sequencing length is greater than the size of the inserts, you will sequence into the adapter on the opposite side.
To determine how frequently this will occur, check the sequencing length relative to the average insert size.
If the fragment size is 150 and the sequencing length is 50, it's highly unlikely that you will sequence into the adapter on the opposite side.
You don't mention the fragment size, but you do specify that the sequencing length is 101 bp, so it's quite possible you sequenced into the adapter on the other side for a portion of your reads. How far you sequenced into the adapter depends on the length of the insert, which varies. During library preparation, the size of the inserts are distributed around the average desired insert size, following a normal curve I believe.

adapter - insert - adapter 
     ->            <-

That's the best explanation I could come up with.
Someone else might do better.

The diagram is meant to illustrate with the arrows that the sequencing starts at the adapters, occasionally goes through the insert, and into the adapter on the other side. The arrows keep moving around, but they should be under the adapters. The diagram illustrates paired-end sequencing. For single-end, just keep one arrow.

[ea11]

Thanks for the reply. That makes sense and does explain what I was struggling with. With regards to fragment size, I am unsure as I didn't do the sequencing, I have been tasked with the analysis of the sequencing.

So does that mean the files should perform fine downstream and align well, despite the reads not all being the same length after trimming? I will be trimming both forward and reverse reads of a particular sample with the same cutadapt options so I'm assuming that should affect alignment?

Thanks

[blancha]

So does that mean the files should perform fine downstream and align well, despite the reads not all being the same length after trimming?

Yes.

I will be trimming both forward and reverse reads of a particular sample with the same cutadapt options

Don't think this is correct. Depends what you mean by the same options.
You should probably just use Trimmomatic which will process both forward and reverse reads correctly in one command. Trimmomatic also already comes with the adapter sequences.

[ea11]

Sorry, yea that was incorrect, the sequences will be different. I will look into Trimmomatic,

Thanks for your help

[GenoMax]

If you want to keep things simple also look at BBDuk.sh from BBMap. I find the options easier to understand, you might too.