Hi everybody,

I'm a bit confused about the automatic model selection. Is there actually anywhere a list, where you can find out what parameters each model string include and which models are the same?
As in another post, I'm a bit puzzled that Model identifies let's say 010232 as the best model, but after entering the parameters in GARD it says at the end of the analysis that it used 10343.
I'm looking for recombination and selection in differenet strains of a virus. When I chose the REV and gamma (idenitified by Modeltest) it found 1 significant breakpoint in GARD and SBP and splitting the sequences and looking at trees in Mega, confirmed at least 2 jumps. However, when using the suggested Model-string or simplot can't detect recombination. So, I would like to have a closer look at the diffrences in the models used.

Thanks!
Marlene

Hi Marlene,

If you look at page 16 of Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login, you can find an explanation of what 6-character model strings encode. 010232 and 10343 (it really should be 010343 - I'll fix the bug in datamonkey that cuts of the leading 0) are the same two models.

Cheers,
Sergei

Was just about going to take my question back. Thank you anyway!
And maybe for other users: the code describes how many parameters are used to describe rates in the following order:
a <-> c, a <->g, a <-> t,  c <-> g, c <-> t and g <-> t

Hi Marlene,

Yep, you got it  :)
The original six letter notation goes back to about 1995 in PAUP*

Best,
Sergei