Hi Felipe,
You didn't tell us what function you used: I assumed it was ace() in ape.
Often, the fact that SEs cannot be computed by ace() means that the
model is a poor fit and that a simpler model is better. You said that
the LRT is significant (did you use anova() on the outputs of ace?), so
maybe there is an intermediate model between ARD and ER that fits better.
Best,
Emmanuel
Le 29/02/2016 19:25, Felipe Rossetto a écrit :
Hello everyone,
I am estimating the ancestral state for a 3-state discrete character. The
log-likelihood test indicated the all rates different model (ARD) fits well
for this character, however, thre was message warning NANs were produced.
So, I looked for standard error typing the command:
ARDreconstruction$se
and only NaNs were detected:
[1] NaN NaN NaN NaN NaN NaN
Then, that means there is not confidence in the recosntructed rates using
ARD model? Then is it better ER for estimating ancestral state for this
character?
Many thanks
Felipe Rossetto
Herbário FUEL
Departamento de Biologia Animal e Vegetal
Universidade Estadual de Londrina
Rodovia Celso Garcia Cid, PR 445 Km 380, Campus Universitário
Caixa postal 10.011
CEP: 86057-970
Londrina, Paraná, Brasil
CV: http://buscatextual.cnpq.br/buscatextual/visualizacv.do?id=K4211230P7
[[alternative HTML version deleted]]
_______________________________________________
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
_______________________________________________
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
