Sequence Trimming without Knowing Primer Sequences

[KyuzoSeq]

Hi,

I have an IonTorrent output but the provider doesn't want to give information about primer sequences. Instead it gives a bed file includes regions of interest. So if I do alignment and bedtools intersect than trimming sequences by outside parts which is not in the region, I'm not comfortable this is a good approach. We want to use the output as a data for finding SNP frequencies and for the best approach primer sequences must be trimmed.

So, is there a way to do this job properly? I mean is there a way to trim sequences with no knowledge about primers?

[Brian Bushnell]

What kind of experiment are you doing? Is it some kind of exon-capture, or an amplicon approach? Or, why do you expect to see primers in the sequence?

[KyuzoSeq]

What kind of experiment are you doing? Is it some kind of exon-capture, or an amplicon approach? Or, why do you expect to see primers in the sequence?

It's basically a variant analysis process. What I need is trimming sequences by primer sequences but we have no information a about primer sequences.

[Markiyan]

Use fastq contaminant detection and removal approaches. Do a bit of de novo...

If you would like to have some idea about the primers you can try:
Converting to fastq format and having a look at fastqc at high kmer settings (-k8 or -k10).

Up the memory to java wm (-Xmx) if you run out of memory for the fastqc application.

Also try assembling a susbset (1k - 10k) of reads matching the primer-like kmers, and having a look at the de novo assembly results in consed or similar program.
Newbler assembler has primer detection capability and may give you some idea about the primers present (check it's log).

Local setup of WWWblast / STARblast or similar tool comes very handy for doing investigations.

grep -c primer *.fasta against the sequences in fasta format may give you an idea about the abundance of the given kmer.

Other options to investigate includes trimmomatic and similar tools.

Also look for high quality unaligned region a the the ends of your reads when examining the mapping results.