#12 per SNP Tajima

[Anonymous]

Originally created by: EricSu... (code.google.com)@gmail.com

When I calculate Tajima's D on a per nucleotide basis (--window-size 1, --step-size 2) and average the Tajima's D for sites within a particular gene, I get a different number than when I use Variance-at-position for that gene, even though it uses the same nucleotide positions.

I would expect this number to be the same, are there any additional calculations behind Variance-at-position.pl? Why is this value not a simple average of the per-SNP average for that gene?

[Anonymous]

Originally posted by: xuanyuli... (code.google.com)@gmail.com

I am also confused about it