Hi Andrew,

Branch-site REL does not directly test for purifying selection, so you can't extract the information about it from the results table. You can get an indication of what the selection profile looks like, but there is no formal test. You could readily modify the test to ask questions about dN/dS < 1 (i.e. is there a proportion of sites along a branch where dN/dS), but I think for the vast majority of branches in the vast majority of data sets the answer is yes by default, i.e. one will have to work hard to find branches where no sites evolve with dN/dS < 1.

Can I ask why you wish to run this test?

Sergei