High Kmer Content in middle of the read

Question

[Sam]

Hi there, we have now got the 12 sequencing samples back. However, when reading the FastQC report, we saw something weird: A large kmer content spike in the middle of the read.

There is slight amount of adapter leftover but they should inflate the kmer content at the start of the read. Has anyone seen something like that before? Is it something to be worried about?

The full report for a pair of the reads can be found here. Thank you very much!

Answer 1

[Ian]

I was told that this represents mis-incorporation of Truseq adapters into the sequence. But in my experience this only effects a minority of reads; tens of thousands out of millions. Sorry, no reference for this.

Answer 2

[harold.smith.tarheel]

The over-represented sequences in your kmer reports (TATCTCGTATGCCGT and GTGGTCGCCGTATCAT) are perfect matches to segments of the TruSeq Index and Universal adapter primers, respectively. Without more details, it's difficult to know why only a portion of the adapters are present, and only at the positions indicated.

Answer 3

[garcesj]

Hi @Sam, I've got exactly the same problem with my samples... can you finally solve it?? I'd expect that adapter contamination or mis-incorporation were in the 5'/3'-ends but I don't understand why it's appearing in the middle...

Thanks a lot!

[Sam]

Unfortunately, that's like 3 years ago and I have no memory of that. I think @Ian's answer should most likely be what's happening

[garcesj]

Ok, thanks anyway... so I guess, according to Ian's response, that's useless try to do some kind of trimming, isn't it? Thanks again!

[Sam]

Maybe just check the number of reads that show this kind of problem. If the number is small, just remove them