I have used the model "2 or more dN/dS ratios for branches" in the PAML analysis. However, for some branches, the ka/ks ratio (omega value) is 999.0 (ks is too small). I have checked the alignment and I found nothing wrong there. Should I ignore these branches when interpreting the result? Any advice would be appreciated!
If you do a proper Google search, you would find an answer from Ziheng Yang himself:
https://groups.google.com/forum/#!searchin/pamlsoftware/999 and https://groups.google.com/forum/#!searchin/pamlsoftware/999/pamlsoftware/Lq7nYfTdZb0/57d7kEIi6dkJ This sort of question is better posted on software specific forums where the developers are at hand to give more specific answers. They know their programs inside out.