MNM — Rate matrix on triplets/codons that includes
multi-nucleotide mutations
Usage
MNM(submodel: CTMC<n>, v2: Double, v3: Double, a: t)
→ CTMC<t>
Arguments
A default beginning with ~ specifies a prior
distribution.
-
submodel: -
Nucleotide rates
-
v2: -
Relative rate of 2-nucleotide mutations
-
Default:
~LogLaplace(-9,1) -
v3: -
Relative rate of 3-nucleotide mutations
-
Default:
~LogLaplace(-12,1) -
a: -
The Triplets alphabet
-
Default: The alphabet in the current context
Original default expressions
-
a -
get_state(alphabet)
Description
Allow mutations to change one, two, or three nucleotides within a codon or triplet. Single-nucleotide rates come from the input rate matrix. Two- and three-nucleotide mutation rates are proportional to products of target nucleotide equilibrium frequencies at the changed positions, without a transition/transversion preference.
For reversible inputs, v2 and v3 are the equilibrium rates of two- and three-nucleotide mutations relative to single-nucleotide mutations. The fraction of events that change two nucleotides is:
v2/(1+v2+v3)
The fraction of changed nucleotides attributable to these events is:
2*v2/(1+2*v2+3*v3)
These ratios describe the mutation model; subsequent selection modifiers can change them.
For non-reversible inputs, the same ratios hold under a reference distribution: multiply nucleotide equilibrium frequencies across the three positions and normalize over allowed codons or triplets. Actual equilibrium ratios can differ.
Stationary inputs use the equilibrium distribution of the constructed rate matrix at the root. Non-equilibrium inputs instead use root frequencies calculated from nucleotide root frequencies, independently of v2 and v3.
Examples
GTR +> MNM +> dNdS|w: GTR +> MNM +> dNdS(w)| +> M3