Codeml giving different chi2 values when reducing the number of sequences

Hi,

I am using codeml for my positive selection analysis. I have 93 E. coli sequences for my analysis. I run the analysis using the 93 sequences against NSsites model 1 and 2 and the corresponding chi2 value after analysis was around 25 which means it was under positive selection. However, I later divided this analysis into two parts where one alignment file had 40 sequences and other had remaining 53 sequences and now when I run the analysis using codeml, then the individual chi2 values are 0 for both analysis. With more samples, the chi2 value was significant but now with the same sequences and just the number of sequences are reduced, the chi2 value is 0. If possible, can anybody please explain me why it is happening like this??

• 12 views

Read more here: Source link