Amplicon sequencing and unwanted amplicons

Question

[misterie]

Hi,

I have sequenced amplicons of genes on different chromosomes in order to call variants. I noticed, that many sequences were mapped to another region in the genome –– on a chromosome that should contain only one amplicon, but I get variants for two genes that overlapped different regions of this chromosome. How is it possible? Is it a problem with my pipeline or is it be caused during sequencing?

I mapped amplicon reads against the whole reference genome using bwa-mem. I called variants using GATK.

[GenoMax]

Did you check your primers against the entire genome to make sure there only predicted products were possible i.e. primers were uniquely amplifying region you wanted? This is generally done using blat or primer blast. You can still do this and see if multiple locations in the genome are likely to be amplified.

Since you have already done the experiment, if you are sure that there should be only one valid product per primer pair then you could try to make alignment parameters more stringent to see if you can eliminate some of the secondary alignments.

[misterie]

Yes, I did. Primers are uniquely amplifying region. I see, that the mapping quality in these regions is very high. Do you have any other ideas?

[GenoMax]

Primers are uniquely amplifying region.

If that is the case then why are you getting

I noticed, that many sequences were mapped to another region in the genome

What does the following mean?

a chromosome that should contain only one amplicon, but I get variants for two genes that overlapped different regions of this chromosome

You are getting reads aligning to two different genes on same chromosome? Are they aligning to a domain that is shared by the two?

[misterie]

I mean that variants called on one chromosome are localized in two totally apart genes. Only amplicon from one of these two genes was sequenced.