[R-sig-phylo] error message using compar.gee

Emmanuel Paradis emmanuel.paradis at ird.fr
Mon Oct 10 15:19:53 CEST 2011


Hi Kaspar,

The error comes certainly from that species with "ones" are very closely related (or a similar pattern) making impossible to estimate correctly the binomial parameters. You may try transforming the branch lengths of the tree to check this.

About the warning, have you tried:
 
all(rownames(birds1) %in% tree2$tip.label)

?

Cheers,

Emmanuel

-----Original Message-----
From: Kaspar Delhey <kaspar.delhey at monash.edu>
Sender: r-sig-phylo-bounces at r-project.org
Date: Mon, 10 Oct 2011 22:09:36 
To: <r-sig-phylo at r-project.org>
Subject: [R-sig-phylo] error message using compar.gee

Hi list members,

I would be very grateful for input to solve the following problem.
I am trying to run a comparative analysis using compar.gee from the 
package "ape".

The dependent variable is categorical and can take values of 0 or 1 (317 
"zeros" and 55 "ones"). Thus I fitted a model using family=binomial.

I get the following error message which seems to point at two problems:

>  compar.gee(col.var~range+habitats,data=birds1, family=binomial, phy=tree2)
Beginning Cgee S-function, @(#) geeformula.q 4.13 98/01/27
running glm to get initial regression estimate
  (Intercept)        range     habitats
-2.246230008  0.003379506 -0.045287365
Error in gee(col.var ~ range + habitats, c(1, 1, 1, 1, 1, 1, 1, 1, 1,  :
   Cgee: error: logistic model for probability has fitted value very close to 1.
estimates diverging; iteration terminated.
In addition:Warning message:
In compar.gee(col.var ~ range + habitats, data = birds1, family = binomial,  :
   the rownames of the data.frame and the tip labels of the tree do not match: the former were ignored in the analysis

>
	



The first problem seems to be a failure to converge and I am not sure 
whether it has a solution. Note that if I use any other family (i.e. 
gaussian or poisson) the model produces meaningful (albeit wrong due to 
incorrect error structure) estimates. I would be most grateful for any 
suggestion.

The other warning message seems to indicate that the data does not match 
the names on the phylogeny. I have checked this and I see no mismatch. 
Moreover I have sorted the data in such a way that the order is the 
same. Again, any suggestions welcome.

Thanks in advance and best wishes

Kaspar

-- 
Kaspar Delhey
e-mail: kaspar.delhey at monash.edu
Bldg.18 School of Biological Sciences
Monash University
Clayton, 3800 Victoria
Australia


	[[alternative HTML version deleted]]

_______________________________________________
R-sig-phylo mailing list
R-sig-phylo at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo


More information about the R-sig-phylo mailing list