Hey again,

it looks like vcf-kit is more straightforward for calculating Tajima's D value in sliding windows! To be honest, I haven't been really able to make PopGenome function, but vcf-kit is quite easy:

vk tajima 5000 3000 input.vcf > output.TajimaD

5000 refers to the window size and 3000 the step size of the sliding window.

I am not aware of a way to do this; however, you could simply run --TajimaD multiple times with varying bin size and then find some way to provide a summary function. For example, if you converted your outputs to BED format, you could likely then use BEDTools to summarise the data into a single output. This is experimental, of course.