2 questions: FcLM with 1.6.12 and comparing AICc of codon version DNA models (partitioned)

[Karen Meusemann]

Hi all,

I have 2 quick questions:

1) referring to the bug in 1.6.12 with -spp and FcLM: could this bug be fixed already (1.6.9 ran smoothly)

2) I have a small alignment (nuc protein coding) - now trying to find the best model for a) partitioned (1. codon pos, 2. codon pos, 3. codon position) and b) using a codon model (only one marker): Can I actually compare here the AICc or isn't this reasonable since the parameter space is probably completely different?

Many thanks & best

Karen

[Minh Bui]

Hi Karen,

On 30 Nov 2019, at 10:08 pm, 'Karen Meusemann' via IQ-TREE <iqt...@googlegroups.com> wrote:

Hi all,

I have 2 quick questions:

1) referring to the bug in 1.6.12 with -spp and FcLM: could this bug be fixed already (1.6.9 ran smoothly)

I haven’t had enough time to inspect this issue further. But will get to it next week.

2) I have a small alignment (nuc protein coding) - now trying to find the best model for a) partitioned (1. codon pos, 2. codon pos, 3. codon position) and b) using a codon model (only one marker): Can I actually compare here the AICc or isn't this reasonable since the parameter space is probably completely different?

The AIC/BIC scores between DNA and codon models are directly comparable. So you can safely compare these scores from .iqtree file.

However, they are not comparable with a protein model, one needs to perform some adjustment to the likelihoods and degrees of freedom. See Simon Whelan paper (https://www.ncbi.nlm.nih.gov/pubmed/25209223). This method is now implemented in IQ-TREE version 2, with the --modelomatic option. 

Cheers,

Minh

Many thanks & best

Karen

--
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/a62cf6ae-e48e-4a89-af5a-2825bbffb9f6%40googlegroups.com.