Miseq:Trimming, and sequencing primers at the beginning of a read

[clintp]

I noticed that when I trim my Miseq reads for adapter contamination (getting rid of the 3' portion of the read), I could still grep the trimmed reads for ACACTCTTTCCCTACACGAC (the sequencing primer/adapter sequence) and find several thousand at the 5' end of Read1 reads. These shouldn't be there, right? What am I missing?

I used fastq-mcf to trim the 13bp common TruSeq sequence AGATCGGAAGAGC.

Primer sequences do not appear in the beginning of Read2 reads. In the sample sheet, I did not request that the MiSeq do any onboard trimming. For library prep, I used NEBNext Ultra, whose adapters, seq primers, and indicies are the same as the TruSeq stuff.

So, my questions are 1) why am I getting primer sequences in read 1? and 2) Is the 13bp sequence sufficient for trimming Illumina reads (and should I be doing this differently--the reads are used for de novo assembly and blast-based binning, so aggressively getting rid of adapter sequences is important to me)?

[GenoMax]

First part could be explained by having adapter/primer dimers without any insert.

As for trimming give "trimmomatic" (http://www.usadellab.org/cms/?page=trimmomatic) or cutadapt (http://code.google.com/p/cutadapt/)/trimgalore (http://www.bioinformatics.babraham.a...s/trim_galore/) a try. Recent comparison of trimmers available http://www.plosone.org/article/info:...l.pone.0085024.

[clintp]

I like the idea of trimmomatic, but I can't seem to make it trim the adapters--they still show up after the following:

java -classpath /opt/Trimmomatic-0.32/trimmomatic-0.32.jar org.usadellab.trimmomatic.TrimmomaticPE -threads 8 -trimlog TT.log Pool1_S1_L001_R1_001.fastq Pool1_S1_L001_R2_001.fastq p1r1_TT.fastq p1r1_To.fastq p1r2_TT.fastq p1r2_To.fastq LEADING:3 TRAILING:3 ILLUMINACLIP:adapter_13.fa:2:30:10 SLIDINGWINDOW:4:15 MINLEN:16

I may have the parameters set funny, but I don't know the best way to set it. My adapter sequence is the 13bp common Illumina sequence--is 13bp not scoring high enough to get trimmed?

[GenoMax]

Are you using the raw data for trimming? Why not use the TruSeq3 (PE) adapters that Trimmomatic includes (you will find those files in "Trimmomatic-0.30/adapters/") for the ILLUMINACLIP input.

[mastal]

@clintp

I think the parameters you are using for the IlluminaClip step (2:30:10 ) are too high for trimmomatic to recognize a match to a 13-base adapter sequence;

You need to either change the values or use a longer adapter sequence.

See the trimmomatic web page,

http://www.usadellab.org/cms/?page=trimmomatic

particularly the discussion under the heading 'Adapter Fasta', from which I have extracted this quote:

'The thresholds used are a simplified log-likelihood approach. Each matching base adds just over 0.6, while each mismatch reduces the alignment score by Q/10. Therefore, a perfect match of a 12 base sequence will score just over 7"

[clintp]

@mastal
Yep, understanding the cutoff scores helped a lot (durrr). Somehow I missed that discussion on the trimmomatic page.

@GenoMax
Thanks for that reference--very useful. It's too bad they didn't include ea-utils/FastqMcf in that analysis, though.