Getting SNPs that are the complementary base of the reference

Question

[Matteo Schiavinato]

Hi there,

maybe a dumb question but it's the first time this happens to me. I am doing variant calling with mpileup and bcftools call, and while assessing the results in a browser I stumbled across some variants that were, as a matter of fact, originating from reads that were representing only the complementary base of the reference.

Practical example:

I have 1 position which is T, its complementary is therefore A. All my sequencing reads show an A in that position: I would expect a calling algorithm to NOT call that position as variant, as it is just the complementary base. However, the algorithm calls it. How can I avoid this?

[WouterDeCoster]

Could you add an IGV screenshot to illustrate your issue? I am a bit confused. Sounds like a normal T/A SNP to me.

[Pierre Lindenbaum]

I have 1 position which is T, its complementary is therefore A. All my sequencing reads show an A in that position:

There IS a mutation T->A at this position.

[Matteo Schiavinato]

@wouter @pierre thanks for the answers, I sorted it out after some plotting. All the reads mapping on the forward strand (which is T) had an A, so the whole read maps on the strand which should have a T at that position but has an A instead. Hence the SNP. Thanks anyway!