Table 2 Open in new tab The way that...

Table 2

The way that different MSS models model variation among synonymous substitution rates, $α (x, y)$

Model	Functional form	Parameters in addition to basic MG94
Standard codon	$α (x, y) := 1$	0
A priori	$α (x, y) = α_{i}, i = 1.. K$	K−1, 1 < K < N, where N is the number of amino acids that have synonymous codons that can be exchanged by a single nucleotide substitution, e.g. K = 2 in Rahman et al.
Genetic algorithm (GA)	$α (x, y) = α_{⟨ i ⟩}, i = 1.. K$	K−1, 1 < K < N. Same as above, except the allocation of specific (x,y) pairs to the ith rate class is not fixed a priori but inferred using a heuristic discrete space search.
SynREV	$α (x, y) = α_{A}$ A is the amino acid encoded by x and y	N−1 (18 for the Universal genetic code)
SynREVCodon	$α (x, y) = α_{x y} = α_{y x}$	P−1, where P is the number of synonymous codon pairs within one nucleotide substitution of one another (67 for the Universal genetic code)

Model	Functional form	Parameters in addition to basic MG94
Standard codon	$α (x, y) := 1$	0
A priori	$α (x, y) = α_{i}, i = 1.. K$	K−1, 1 < K < N, where N is the number of amino acids that have synonymous codons that can be exchanged by a single nucleotide substitution, e.g. K = 2 in Rahman et al.
Genetic algorithm (GA)	$α (x, y) = α_{⟨ i ⟩}, i = 1.. K$	K−1, 1 < K < N. Same as above, except the allocation of specific (x,y) pairs to the ith rate class is not fixed a priori but inferred using a heuristic discrete space search.
SynREV	$α (x, y) = α_{A}$ A is the amino acid encoded by x and y	N−1 (18 for the Universal genetic code)
SynREVCodon	$α (x, y) = α_{x y} = α_{y x}$	P−1, where P is the number of synonymous codon pairs within one nucleotide substitution of one another (67 for the Universal genetic code)