Expected Mutation Rate Matrix

Question

[Arjen Ten Have]

In an attempt to identify new subfamilies I would like to use the matrix that is the basis of PAM and Blosum. That is, the matrix with expected mutation rates. PAM and Blosum are log odds of the real and the expected "mutation" rates and hence are indicators of similarity. But they use averaged values for the real "mutation" rate (this actually should be evolution rate), I have a single value derived from a column of my MSA. Anyone knows where to find this? I could calculate it myself based on the genetic code but it seems likes reinventing the wheel.