Unexpected results using different genome sets

[444thLiao]

Hi, popcogenT is excellent software and we have used it to resolve many questions.

However, one situation we meet is that we have gradually increased our genome set. When the gene set is small (~500 genomes), we resolve them in to different populations. For example, genome A,B,C all belong to pop1. When the genome set is larger (~1000 genomes), we resolve them again. In this time, genome A is pop2, genome B,C are still pop1.

This is the background to my following questions.

1. When I got through your code, I found out that you set alpha=0.1 in the method summary_frame. But within the length_bias.txt, they should be 95% CI. Accordingly, should we set alpha=0.05 ?
2. I believed that my questions came out because of the Negative selection cutoff changes along with the changes of the number in genomes. Is it normal, or how could we deal with it?