Dear R members: Sorry, this question is not tightly related to R software, but I usually use R to do this type thing. So please if you can help me out! I want to predict a protein function, by through sequence alignment, I get several low identity sequences(say 100) in several classes(say 5 different classes), I have the Z-score for every sequence alignment, how to calculate the class prediction reliability? Thanks so much!! regards, Bing