FastQC, Kmer count, Trimmomatic: no success in trimming, still fail Kmer

[skmotay]

I'm using RNAseq for DGE and failed in trying to map paired end reads in TopHat2. Samples passed the first two parts of fastqc, but failed per base and Kmer count. Some samples did pass per base.

I tried trimming them to remove adaptors, but nothing improved the Kmer plot. http://imgur.com/Fhbx534 All of the kmer plots from this data set (18) look similar. I used the adaptor settings built into trimmomatic, but none worked. I then emailed genewhiz and asked for their (truseq) adaptors, but trimming for those sequences didn't improve the plots.

Ex per seq quality score http://imgur.com/5L5PaJU,Hg2mHk0#0
per base http://imgur.com/5L5PaJU,Hg2mHk0#1

Am I doing something wrong? Should I just trim everything across certain bases?

Edit: reads are paired

adaptor input file:
>PrefixPE
GATCGGAAGAGCACACGTCTGAACTCCAGTCAC
>Prefix PE
AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT

[Brian Bushnell]

Are your reads paired? If so, you can trim them with BBDuk via overlap, without specifying the adapter sequence, like this:

bbduk.sh in1=r1.fq in2=r2.fq out1=trimmed1.fq out2=trimmed2.fq tpe tbo

That package comes with both TruSeq and Nextera adapters, which are another possibility that you can try.

[skmotay]

Yes, reads are paired.

I think I should have mentioned that I'm working in Galaxy, which I think limits what tools I can use. I don't see BBDuk in my tools.

[Brian Bushnell]

Ah, too bad. I'll see if I can get BBTools put into Galaxy. Though you can still download it and try out the Nextera adapter file. I would recommend, though, that you contact the source of the data to find out what kind of adapters were actually used - that will save a lot of time.

[skmotay]

I emailed genewhiz and they sent me the adaptor sequences (listed I original post). I'm not quite sure I wrote the adaptor sequence file correctly and I couldn't find a clear example.

[avo]

Does Fastqc give you a warning for overrepresented sequences?

I don't know the default -k setting for fastqc but maybe it helps to increase it.
My impression is that the adapter trimming worked. But the overrepresented k-mers test fails due to other reasons

[skmotay]

They all pass overrepresented sequences.