Allele frequency plot : imputed vs reference panel variants

Question

[hello_genome]

After imputation, I plotted the sample allele frequency (AF) values against the panel AF values (here, I've randomly selected 100,000 variants of chromosome 2 as the data set is too large).

Ideally, there should be a uniform, fairly tight diagonal line, with an increasing slope of 1.

What could be the reasons why I'm not observing this tight diagonal?

[BlastedBadger]

I won't be able to help you, but to clarify the question, are you asking about the dots that are stacked near the axis lines? Could be that the program (which one did you use) is not reliable for low AF? Did you filter for a minimum allele frequency in both panels?