NGS Adapter Sequence Cleaning

[Mehmet Göktay]

Hi everyone,

I have miseq data which needs to be cleaned. I also know the sequence of the adapter.
The problem is when I used cutadapt to remove the adapter sequences from my reads, most of the reads are cleaned but when I blast adapter sequence against all cleaned reads I see pretty small portion of my data contains adapter sequence. I don't get it. Even though I used the parameter which allows me to clean adapter sequence anywhere in the read.

May be cutadapt is now the adapter cleaning program which I'm looking for. Is there any suggestion for me which can include other adapter sequence cleaning programs or using other parameters which will make me get rid of the remaning adapter contamination.

Best Regards

[dpryan]

Try looking at some of the blast hits and comparing the sequence to that from before trimming. Especially if the hit is near the end and is only partial, then earlier adapter trimming from a read could create false apparent adapter matches (since if it's partial and sequence was already trimmed from it, then it's not real adapter sequence).

[GenoMax]

Give BBDuk a try after following Devon's suggestion above.

[Mehmet Göktay]

Firstly thank you for your kind answers.

I have already looked to the cleaned sequences and found out that part of the adapter sequence is located almost at the end of the 3' end. But most probably beacuse of the fragment size which is too small sequencing machine sequences the fragment following adapter sequence and because of the read size is 300 nucleotide after adapter sequence it just adds random nucleotides until it reaches to the 300 nucleotide.

As I understand cutadapt can not handle this situation eventhough i used all available parameters for location of the contaminant.

[dpryan]

Depends on the sequence you told it to trim. In general, it can handle that. However, perhaps you'll find BBDuk easier to use.

[Brian Bushnell]

Also, if you are using paired reads, you can give BBDuk the "tbo" flag (trimbyoverlap) in addition to the adapter sequences. This will allow trimming of adapter sequences at the very end of the read, which are too short to detect by sequence matching.

[bastianwur]

Don't forget that you can change the accepted error rate in cutadapt. Some of the adapters might just have one base more than the error rate, then it'll not be trimmed.

Unrelated: Cutadapt also only checks for the given sequence, not for the complement or reverse complement.
Yes, I know, having adapters like this shouldn't happen, but I had reverse complemented adapters in one of my datasets (don't ask me how).

[Brian Bushnell]

Unrelated: Cutadapt also only checks for the given sequence, not for the complement or reverse complement.
Yes, I know, having adapters like this shouldn't happen, but I had reverse complemented adapters in one of my datasets (don't ask me how).

Just FYI, BBDuk does look for reverse-complements, unless you explicitly tell it not to.