What is the convention for CNV vcf output in chromosome X and Y?

Question

[14134125465346445]

What is the convention for a CNV caller that produces vcf output in assigning copy number to the sex chromosomes?

If there is only one copy of chrX and one copy of chrY, is the caller supposed to produce two entries with CN=1 for each of them? Or no entries?

For example:

A human male sample (hemizygous), has no extra chromosomal copies: should the CNV vcf file be empty, as there are no copy changes from the normal? Should it contain CN=1 for chrX and CN=1 for chrY?

A human sample with XXY: should the CNV vcf file report CN=2 for chrX? How is this annotation distinguished from a non-aberrant chrX CN=2?

Answer 1

[Alex Paciorkowski]

Any CNV caller dealing with human data should adhere to the International System for Human Cytogenetic Nomenclature (ISCN). Recent reference: https://www.karger.com/Article/FullText/353118

This system served well for the transition from karyotyping to microarray, and should do the same for next-gen sequencing when vcf's are generated.

In regards your 2 examples/questions:

1. Does the caller know a priori the sex of the sample, and therefore the expected ratio of X/Y copy number? Or are you using the caller to determine sex of the sample? If it's the former, then in theory the vcf for the X/Y chromosomes could be empty (although unlikely with real data since it is quite normal to harbor benign CNVs on X/Y...). If it's the latter, then the X/Y files would show CN=1 for the loci evaluated -- and would be a tip-off that the sample is XY.

2. A human sample with XXY would be (X) * 2,(Y) * 1. This output would be different than XX = (X)*2 because copy number for chrY=0 in that case.