Why is the AFS returning decimal values?

[MAlejandraJimenezM]

Hi, im new to tskit and I am using it to generate the Allele Frequency Spectrum (AFS) with the following R code:

afs <- ts$allele_frequency_spectrum(sample_sets = NULL,
                                    windows = NULL,
                                    mode = "site",
                                    span_normalise = FALSE,
                                    polarised = FALSE)
print(afs)

Nothing crazy!

However, when I inspect the output, I notice that some values are decimals instead of integers:
[1] 0.0 47.0 25.5 18.0 12.0 4.0 8.0 3.0 3.0 4.0 6.5 ...

I expected the AFS values to be integers since they represent allele counts at different frequencies.

What could be causing the presence of decimal values in the AFS? Is this expected behavior due to how tskit calculates the spectrum? normalization? or is there an additional step I should take to obtain only integer values?

I’d appreciate any insights!

[tforest]

Hi!
I guess it's because you have sites at which there are more than 2 alleles.

Take a look at https://tskit.dev/tskit/docs/stable/stats.html#allele-frequency-spectrum, especially paragraphs 2 and 4.

"if we are computing the unpolarised AFS, we add one half to each entry of the folded output corresponding to each allele count including the ancestral allele"