Recommendations and warning
Most population genetic analyses rely on accurate estimates of real
populational genotype frequencies, including here
ClonEstiMatePoly method. The number of different possible
genotypes at one locus increases with the ploidy and the number of
alleles (Reichel et al. 2015). We thus draw users’ attention on the fact
that sample sizes should naturally be larger in polyploid organisms to
accurately estimate their genotype frequencies, despite the fact that
genotyping more alleles per individual may help assessing allele
frequencies.
Missing values and null alleles compromise comparisons between
individuals, lineages and populations, and are susceptible to create
biases and misinterpretations. Suspected null allele can be coded as
unknow allele with their own specific letters or positive integers, and
should be clearly reported before interpretations. Indeed, no
“correction” or “assumptions” can enhance a blurred and incomplete
genotyping signals without deep consequences on the computed indices and
then their interpretations, whatever the ‘correction’. We thus recommend
users to rather remove genetic markers and individuals with missing
values and uncertain genotypes.