Certain chromosomes differ from the overall ploidy of an organism (e.g. sex chromosomes). Hence, variants occuring at such chromosomes exhibit different allele frequencies. This can best be modeled via a species definition, see e.g. this scenario. In case the required prior knowledge about mutation rates and heterozygosity is unknown, it is altenatively possible to constrain allele frequency universes by chromosome. The following example defines the calling for two human samples under a uniform allele frequency universe.
As can be seen, we define a universe for all chromosomes (all), and add exceptions for Romeo (the male) being haploid in the sex chromosomes.
For Juliet (female), the X chromosome is diploid like all other chromosomes, while there is no Y chromosome at all.
Hence, the universe for the Y chromosome becomes 0.0 only.
Currently, Varlociraptor assumes a uniform prior over the defined universes (the grammar will be soon extended to allow other priors as well).
This means that restricting the universe as above also impacts the calculated probabilities for the defined events.
For example, on chromosome Y, there will be no calls for Juliet, because all allele frequencies other than 0.0 will get a probability of zero.
Therefore, all chromosome Y variants will be classified as romeo_only or absent (the implicit event that none of the samples hosts the variant).
samples:
romeo:
universe:
all: "0.0 | 0.5 | 1.0"
X: "0.0 | 1.0"
Y: "0.0 | 1.0"
juliet:
universe:
all: "0.0 | 0.5 | 1.0"
Y: "0.0"
events:
both_het: "juliet:0.5 & romeo:0.5"
both_hom: "juliet:1.0 & romeo:1.0"
juliet_only: "(juliet:0.5 | juliet:1.0) & romeo:0.0"
romeo_only: "(romeo:0.5 | romeo:1.0) & juliet:0.0"