Pairwise distance matrix with nucleotide substitution model

Question

[luca.zoccarato.88]

Dear community,

I have a sequence alignment (full 16S rDNA) and I'd like to calculate the sequence pairwise distance matrix using a nucleotide substitution model. The task could be easily achieved with the dist.dna() function of the R'package ape but there are not a lot of evolutionary model included (at least not the one I need). Indeed, screening my alignment with jModelTest I got that the best nucleotide substitution model for my data is TVM+I+G (invariable sites and rate variation among sites, respectively).

Does anyone know how I can get a distance matrix computed using this (TVM+I+G) substitution model?

Many thanks!

Luca

Answer 1

[luca.zoccarato.88]

I eventually find out the solution. I report it in case someone might need it. jModelTest provides for the best solution the parameters (e.g. base frequency, gamma and proportion of invariable sites) to perform the related maximum likelihood analysis in PAUP (now available for free at http://paup.phylosolutions.com/). The parameters have to be specified when calling "Lset".

In my case the commands in PAUP were:

Executemy_align.nex
Lset base=(0.2167 0.2409 0.2804) nst=6 rmat=(1.0169 3.1065 1.6336 0.9465 3.1065) rates=gamma shape=0.5980 ncat=4 pinvar=0.1260
Dset distance=ml