Most accurate way to measure adapters quantity in NGS illumina reads

Question

[Denis]

I need to make benchmark of several tools for adapters removal (like trimmomatic, cutadapt, etc.) on my 18S MiSeq metagenomics data. I'm wondering how i can measure the quantity of survived adapters in my data after trimming (so called False Negatives)? I'm thinking about usage of blastn-short option of standalone blast for that. Is it correct and most effective way to do the job. Any suggestions would be appreciated. Many thanks, Denis

Answer 1

[chen]

What's your read length? Is it pair-end or single-end?

If your sequencing is pair-end, like 2*150, for those reads from DNA template shorter than 150, then you may find adapter bases from the tails of both read1 and read2.

How to find these adapters? There are 3 different overlapping patterns for different DNA template length (TLEN):

1, not overlapped TLEN > 2x150

ATCGATTTAGTTT...ATTAGGGATTA
------------------------------------------------------------TGTAATCGTAGT...AATACGATCGA

2, overlapped with 150 < TLEN < 2x150

ATCGATTTAGTTT...ATTAGGGATTA
-------------------...-----------AGGGATTACTATCT...AGATTC

3, overlapped with TLEN < 150

ATCGATTTAGTTT...ATTAGGGATTA-adapter
ATCGATTTAGTTT...ATTAGGGATTA-adapter

Try to search for pattern 3, and count the adapter bases. It should be close to 0, if the data is really clean.

BTW, I also developed a tool AfterQC(http://github.com/OpenGene/AfterQC, Automatic Filtering, Trimming, Error Removing and Quality Control for fastq data) to automatically cut the adapters by utilizing the pattern 3.

[Denis]

Thanks for your response! I apologize not all experiment design details were provided in initial post. So we have 250X250 MiSeq PE reads. As to TLEN in my experiment, it corresponds rather to pattern 2 than 3. Which tool i should use to search adapters in clean data? Is blastn short Ok for that?

[chen]

You'd better program by yourself, it is not difficult.

You can take a look at code of AfterQC and may find your way.