Degree of freedom DTL vs DL

roberto feuda

unread,

Sep 17, 2021, 5:54:29 AM9/17/21

to GeneRax

Hi Benoit,

I am trying to compare the reconciled tree obtained using the DTL vs DL model using the BIC or AIC. But I have a doubt on the degree about the freedom of the DTL vs DL model. My feelings is that for the DL 2 and for the DTL is 3, does this make sense?

Cheers,

Roberto

Benoit Morel

unread,

Sep 17, 2021, 5:59:35 AM9/17/21

to GeneRax

Hi Roberto,

Yes it makes sense. If you have several gene families, it also depends if you optimize global rates (in this case, its 2 and 3) or per-family rates (in this case, its 2n and 3n where n is the number of families).

Cheers,

Benoit

roberto feuda

unread,

Sep 17, 2021, 6:53:54 AM9/17/21

to GeneRax

Hi Benoit,

Thanks for the quick response! I did use the per-species optimization with 172 species (I know it is taking ages), does this affect the degrees of freedom?

Cheers,

roberto

Benoit Morel

unread,

Sep 17, 2021, 7:08:01 AM9/17/21

to GeneRax

Then you need to multiply 2 or 3 by the number of branches in the species tree (2n-2 branches in a rooted species tree with n species)

Le vendredi 17 septembre 2021 à 11:54:29 UTC+2, robert...@gmail.com a écrit :

Reply all

Reply to author

Forward