substitutuion/replacement ratios

[J Hexham]

I am trying to calculate substitution/mutation ratios for a series of human 
autoantibodies. The problem I have is how to score a codon which is mutated 
in all three positions (or in two positions). Clearly this causes an amino 
acid change but is more mutated than a single base change in a codon. Is the
correct strtegy to consider each mutation for the effect it would have had in the absence of any others? Thanx for any help/suggestions. Mark Hexham.