| A |
-u(aπC + bπG + cπT)
|
uaπC |
ubπG |
ucπT |
| C |
ugπA |
-u(gπA + dπG + eπT)
|
udπG |
ueπT |
| G |
uhπA |
ujπC |
-u(hπA + jπC + fπT)
|
ufπT |
| T |
uiπA |
ukπC |
ulπG |
-u(iπA + kπC + lπG)
|
| The table gives rates for any substitution in a nucleic acid sequence or for no substitution at all (the diagonal values).
Base frequencies are given by πA, πC, πG, and πT, the mutation rate by u, and the frequency of change of any base to any other by a, b, c..., l. Rates of substitutions in one direction, i.e., A→G, are generally considered to be the same as that in the reverse direction
so that a = g, b = h, etc. In the JC model these frequencies are all equal, and in the Kimura two-parameter there are only two frequencies, one
for transitions (α) and the other for transversions (β), and the frequency for transitions is twice that for transversions.
PAUP allows these numbers to be varied. This model assumes that changes in a sequence position constitute a Markov process,
with each subsequent change depending only on the current base. Furthermore, the model assumes that each base position has
the same probability of change in any branch of the tree (Swofford et al. 1996).
|
