CNV is usually detected using either read depth or SNP VAF. However, neither of these methods can detect CNV in case of genome doubling (assuming genome doubling is a kind of genome scale CNV comparing to normal tissue which is diplod). We thought of other platform such as SNP array and we did not think they can detect genome doubling as well. Can any one share some comments on the method for such event? Thanks in advance.
1
-
$\begingroup$ Are you talking about SNVs or CNVs? I think you are asking about CNV detection in species that have undergone a while genome duplication event, for example teleost fish. Is that what you mean? Or are you working on a novel genome assembly and want to see if the genome of this species has been duplicated? Please edit your question and clarify. If you give us more detail, we might be able to answer better. $\endgroup$– terdon ♦Nov 7, 2019 at 10:19
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
In a tissue or cell line that used to be diploid and has undergone genome wide doubling, you expect more variants having observed allele frequencies that would also occur in tetraploid organism. That is, observed VAF distribution will have (possibly small) peaks at 0.25 and 0.75, in addition to the diploid peaks at 0, 0.5 and 1. Compare the distributions in the tissue or cell line of interest to those of the control, that is, known diploid. Also if it is a doubled male genome, the sex chromosomes will show a new peak at AF 0.5, that is, heterozygous X or Y variants. This is just an idea, I have never tried to look for such events.
$\begingroup$
$\endgroup$
The following may work, assuming you have an Illumina genome sequence library with decent coverage, preferably amplification-free:
Use kat to produce a kmer spectrum plot of the reads (kmer abundance vs. kmer frequency). When you look at, for example, the 21-mer spectrum of the reads, you would expect to see a peak in the spectrum where the kmer frequency corresponds with the estimated coverage of the genome. For instance, if your reads would be expected to provide 10x coverage of the original diploid genome, then you would expect a peak in the kmer spectrum at a kmer frequency around 10.
If the genome had been doubled, and each of the two copies had undergone a number of mutations, then you would also expect a second peak to be found with half the coverage of the first (at a kmer frequency of 5 in the example). The relationship between the heights of the two peaks would reflect the degree to which diploidisation had occurred post-doubling.
Also, I seem to remember something like a pachytene chromosome spread being used with an optical microscope.
