Hi every body,

I have a basic question about calculating the values for log odds matrix from probability matrix.

Given probability matrix PAM 250 for example, the probability of mutating E to R is 3 (3%). The normalized frequency of R is 0.041. Using the formula 10* log (E to R)/frequency of R, -1 is obtained that is in accordance with the log odd matrix PAM 250. The problem is in the latter matrix the value for E-R is equal to the value for R-E, while when I calculate R to E log odd, -10 is obtained which is not present in the log odd matrix.

My question is whether log odds matrix are unidirectional?

I will appreciate any help

Nazanin