trim_galore and cutadapt to paired-end DNA sequencing

Question

[Shicheng Guo]

Hi my colleagues,

Question 1. trim the adapters when we known them.

Suppose the adapter for my DNA sequencing are as the following:

P5 adaptor: 5' ACACTCTTTCCCTACAC***GACGCTCTTCCGATCT***
P7 adaptor: 5' P-GATCGGAAGAGCGGTTCAGCAGGAATGCCGAG

We can trim the adapter for the raw fastq files with trim_galore or cutadapt as the following:

cutadapt -a AGATCGGAAGAGCA -g GCTCTTCCGATCT -o sample.trim.fastq sample.raw.fastq
trim_galore --paired -a GATCGGAAGAGCA -a2 GCTCTTCCGATCT --retain_unpaired --trim1 S1.read1.fq S1.read2.fq

However, I am little confused that why 610,514 reads containing of "GATCGGAAGAGCA" can be found in read2.fastq?

BTW: GATCGGAAGAGCA is the reverse complementary of GACGCTCTTCCGATCT

Any suggestion?

Question 2. trim the adapters when they are unknown.

Is there any violent and forcible​ method to remove the reads containing the adapters? check each adapter (illumina have hundreds adapters)? because I think these adapters should be not contained by human genome, isn't it? therefore, if the reads contain such adapter sequence, they should be filter out. beat me beat me beat me!!

[Lemire]

BTW: GATCGGAAGAGCA is the reverse complementary of GACGCTCTTCCGATCT

No it's not. That's part of your problem, you're not passing the right sequences as arguments.

Answer 1

[Lemire]

My first comment is that you should upgrade to a recent version of cutadapt, which supports both read1.fq and read2.fd in the command line, along with options -a and -A.

Second, I am not convinced that you should expect to see the sequence GCTCTTCCGATCT in your reads. You have to reverse complement it. Take for example the TruSeq example in http://cutadapt.readthedocs.org/en/stable/guide.html#illumina-truseq

you'll see that the reverse complement of the Universal Adapter is provided with the -A argument. You have to give cutadapt the actual sequences that you expect to see in your reads.

Third, in AGATCGGAAGAGCA, where does the last A come from?

Answer 2

[Brian Bushnell]

So... what's the surprise? And why are you just giving the tool partial adapter sequences instead of complete sequences?