Adapters remains after cutting them out

Question

[agata]

I run standard QA on SHAPEseq reads and MultiFastQC reported the presence of Nextera Transposase Seqeunce (Illumina adapter). I wanted to cut out adapters as usual with trimmomatic, but even though I specified the adapter sequence (and then double-check with minion), adapters are still visible in the post-QA report.

File with adapters:

>adapter_1_FW/1\
 CCGAGCCCACGAGACCTCTCTACATCTCGTATGCCGTCTTCTGCTTG
>adapter_1_RV/2\
ACGCTGCCGACGATCGCATAAGTGTAGATCTCGGTGGTCGCCGTATCATT
>adapter_2_FW/1\
TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG
>adapter_2_RV/2\
GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG
>adapter_3_FW/1\
GCCTCCCTCGCGCCATCAGAGATGTGTATAAGAGACAG
>adapter_3_RV/2\
GCCTTGCCAGCCCGCTCAGAGATGTGTATAAGAGACAG

[GenoMax]

Try bbduk.sh as an alternative. A guide is available.

[agata]

Thank you for suggestion! But the strange thing is that I run also other SHAPEseq reads with the same adapters and they were cut out correctly.

[Carlo Yague]

Can you post the following: (1) the trimmomatic options you used, (2) trimmomatic log file for one sample, (3) the first few lines of the file containing the adapter sequence.

[agata]

Suer - please see the edited post

Answer 1

[Carlo Yague]

Thanks for the complement of information. Everything looks good except that I think you should use the reverse complement of the index sequences. By default, trimmomatic trims the sequence only in forward orientation, but when you sequence into adapter sequence, your read contains the reverse-complement of the index.

[agata]

Thank you! It adding reverse-complement sequences of adapters solved the problem.

[GenoMax]

agata : You can accept @Carlo's comment (which I moved to an answer) to provide closure to this thread.