Q matrix for phylogenetic tree reconstruction [PhyMl]

Question

[Jimpix]

Hi! If I want to create phylogenetic tree using PhyML do I need to have Q matrix as a --aa_rate_file parameter in PhyML? What does it change in the result? And from where I can get this matrix? Does this matrix is related to the model of evolution which I need to choose?

Answer 1

[liorglic]

The Q matrix contains the rate parameters of the various sequence evolution models. In most cases when inferring a phylogeny you don't need to provide the Q matrix, and instead you let the software (PhyML/RaxML/IQTREE etc.) optimize the parameters for you. You actually have three options, or levels of control:

1. Specify the model (e.g. HKY, GTR) and its parameters. This is relevant only if you have some a-priory hypothesis regarding the transition rates.
2. Specify the model but let the software find the ML parameter values for you.
3. Let the software try all models and find the ML model and optimize its parameters.

If you have no specific hypothesis regarding the model and you just want the ML phylogeny, then option 3 is probably a good place to start.

[Jimpix]

"Let the software try all models and find the ML model and optimize its parameters" How to set up this? in PhyML you have to type explicite particular model and parameters or it will be just default, no optimize