Possible Cause for False Homozygous Reads (Heterozygous in Real)

[Nikvailo]

While we sequencing CFTR with long-range PCR + Nextera XT + MiSeq 2x250bp, we detected homozygous variant in sample. (upper part of image)

However, when we sequenced the same sample with short-range PCR with same conditions, the variant seemed heterozygous. (lower part of image)

What could be the reason for that?

Is it possible that our long-range primers bind and amplify only one copy of CFTR?

[Brian Bushnell]

Certainly, if your SNP of interest is linked to another mutation that happens to be at your primer site, you'll get a major bias. Perhaps you should examine both of your long-amplicon primer sites with Sanger.

[ECO]

I also have run into a situation where Illumina's adapter trimming was too promiscuous, and was actually trimming regions of the genome!

Ha! I just found my old example and it was CFTR also (albeit exon 10):

http://seqanswers.com/forums/showthread.php?t=23265

How did you do adapter trimming? Are you able to visualize soft-clipped reads?

[HESmith]

Another possibility is that the exon 3 sequence is duplicated in your genome, and the short-range (but not long-range) PCR primers are amplifying both copies.

[Richard Finney]

Are other exons in your target?
What's the coverage of the other exons relative to exon 3 ?

[Nikvailo]

Thank you for all answers. And sorry for my late reply. We were quite busy.

Certainly, if your SNP of interest is linked to another mutation that happens to be at your primer site, you'll get a major bias. Perhaps you should examine both of your long-amplicon primer sites with Sanger.

Thank you Brian. It is quite possible. We will design primers to check that if there is any variant in the primer binding site.

I also have run into a situation where Illumina's adapter trimming was too promiscuous, and was actually trimming regions of the genome!
Ha! I just found my old example and it was CFTR also (albeit exon 10):

http://seqanswers.com/forums/showthread.php?t=23265

How did you do adapter trimming? Are you able to visualize soft-clipped reads?

We are actually using MiSeq built-in adapter trimming option. I understand you example. I is very strange that, that trimmer can also remove genomic regions. But I couldn't figure out that how such bug results with my problem?

IAnother possibility is that the exon 3 sequence is duplicated in your genome, and the short-range (but not long-range) PCR primers are amplifying both copies.

So you say that one of duplicated exon has variant but other has not? Right?

Are other exons in your target?
What's the coverage of the other exons relative to exon 3 ?

It changes. Because of that the PCR efficiency of each amplicon is different the coverage is different in each of exonic regions.