Help a newbie: adapter filtering

[IMPERATOR]

Hi all,

I'm a virgin here, I was referred to this forum by a wonderful post-doc. Myself, I'm a first year PhD student so go easy on me...

So I'm working with NGS data (duh) and I have a couple of questions.

A little background. I have ~350 fastq paired end read files. Illumina hiseq I'm guessing.

1. Is the adapter sequence the same for each file?

2. How do I determine the adapter sequence? I've been using fastQC and under overrepresented sequences, I get this "GATCGGAAGAGCGTCGTGTAGGGAAAGAGGGTAGATCTCGGTGGTCGCCG"

Now when I run my adapter filtering program, cutadapt for example, is the whole sequence my adapter? The post-doc truncated this sequence to "AGATCGGAAGAGC" is he right?

I'm a little confused on the reasoning behind adapter filtering (I comprehend why you do it; to remove the adapter region which is not a part of the query sequence).

Thanks for all the help.

[GenoMax]

Here is a nice primer on TruSeq adapters: http://tucf-genomics.tufts.edu/docum...q_Adapters.pdf

You could end up with adapter contamination if you have adapter dimers or have smaller than expected inserts. Since they are not part of real sequence you want to remove them (specially if you are going to do any de novo assembly).

Ways of how to filter them:

1. http://www.ark-genomics.org/events-o...-illumina-data

2. http://onetipperday.blogspot.com/201...torprimer.html

3. Trim Galore (http://www.bioinformatics.babraham.a...s/trim_galore/) is a wrapper for cutadapt (mentioned in both blog posts) that makes using cutadapt easy.

4. Entire set of various illumina adapter sequences: http://support.illumina.com/download...es_letter.ilmn

[IMPERATOR]

Quick question, I'm working with a viral sequence ~10kb. Should I be less promiscuous with the filtering because of the small size?

On the other hand, the coverage is quite good ~10,000x coverage. (DEEP SEQ HOMIE)

[GenoMax]

What is the aim of your experiment? Sometimes having too deep a coverage may actually be detrimental e.g. if you were trying to call SNV's.

What % of your reads contain adapters?

[IMPERATOR]

In a nutshell, we want to compare the sequences of virus in people that had successful treatment vs those who have failed.

Eventually we would want to refine our techniques, both in vitro and in silico, in order to capture the viral population within each patient.