removing adapters from illumina miseq paired end output files

Question

[Elizabeth]

Dear all, I am working on assembling the genome of a eukaryotic organism (smaller genome size ~65 Mb). As the first step, I performed a QC on the raw reads file (fastq files recieved after a Miseq 2x 300 run). In the results, I saw that the reads have adapters in the 3' end, also some of the reads are low quality reads, so as the next step, I used skewer to trim the reads (to remove the adapter sequence). The problem is, when I perform a second set of QC on the output files (two output files O1 and O2) from skewer, the results indicate that only O1 has been cleaned of the adapters and the low quality reads. For O2, the low quality reads have been removed, but the adapters are still there. Both these files are basically two files generated for the same organism (since it's a paired end sequencing).

The parameters that I have used to run is

skewer -x GATCGGAAGAGCACACGTCTGAACTCCAGTCACGCCAATATCTCGTATGCCGTCTTCTGCTTG -q 30 -n -l 51 -t 16 -m pe -o output filename.fastq

# -x adapter sequence 
# -q trim 3' reads until quality 30 reached
# -n  remove reads with a lot of n
# -l minimum read length allowed, in this case 51
# -t threads
# -m mode of run, in this case paired end (pe)

I am not sure where am I going wrong. Please advice. Has anyone encountered a similar issue?

[GenoMax]

I am not a skewer user but any time you are trimming read data from paired-end reads you should always do the trimming with both of those reads at the same time. I don't see what the exact inputs/outputs are above.

If you are open to trying a different program then you can use bbduk.sh from BBMap suite. Here is a guide for how to use.

[Elizabeth]

I tried using both the reads as well, but the adapters are still present. I will try bbduk. Thanks.

[Elizabeth]

This leads me to another question, do the paired-end reads be concatenated before trimming ?

[GenoMax]

No. With bbduk.sh you will use two inputs and two outputs. in1=inp_R1.fq.gz_R1 in2=inp_R2.fq.gz out1=trim.R1.fq.gz out2=trim.R2.fq.gz. Reason to do this is to make sure that the output data files stay in sync in case one read gets removed from R1/R2 files (or fails trimmed length) then its mate is also removed from other file.

[Elizabeth]

Thank you. bbduk.sh works. The adapters have been removed and I performed some quality filtering on the adapter trimmed reads as well to ensure uniform length distribution. Thanks again

[GenoMax]

Good to know. You may also want to give bbmap.sh (the aligner included in the package) a try.