[R-sig-ME] nlmer error: ...coercing "labelled" to "CsparseMatrix"

[Andrew Beckerman]

Dear list -

R 2.9.1, OSX

I am trying to re-fit an old nlme model of mine using nlmer, but  
getting an error about labels and the CsparseMatrix.

The data are in the following form

 > head(induct)
    clone    id trt ind  ind.adj
1 barney 0.1a1 0.1  30 21.42858
2 barney 0.1a2 0.1  65 46.42858
3 barney 0.1a3 0.1  30 21.42858
4 barney 0.1a4 0.1  60 42.85715
5 barney 0.1a5 0.1  70 50.00001
6 barney 0.1b1 0.1  60 42.85715

The nlme model was fit via the following, which works well:

#grouped data
gd1<-groupedData(ind.adj~trt|clone,data=induct)

# SS starting values
Asym<-50
xmid<-0.25
scale<-1

induction<-nlsList(ind.adj~SSlogis(trt,Asym,xmid,scale)|clone,gd1)
nlc <- nlmeControl(maxIter = 25)
m1<-nlme(induction,random=pdDiag(Asym+xmid+scale~1),control=nlc)

My attempt at fitting with nlmer, ignoring the groupedData object,  
with a single random effect in the Asymptote and with nlme detached, is:

induction.lme4<-nlmer(ind.adj~SSlogis(trt,Asym,xmid,scal)~Asym|clone,
	data=induct,
	start=c(Asym=50,xmid=0.25,scal=1))

Which is resulting in the following error message and traceback.  Any  
insight, or requests for more of the data graciously accepted.

 > induction.lme4<-nlmer(ind.adj~SSlogis(trt,Asym,xmid,scal)~Asym| 
clone,data=induct,start=c(Asym=50,xmid=0.25,scal=1))
Error in as(from, "CsparseMatrix") :
   no method or default for coercing "labelled" to "CsparseMatrix"

 > traceback()
8: stop(gettextf("no method or default for coercing \"%s\" to \"%s\"",
        thisClass, Class), domain = NA)
7: as(from, "CsparseMatrix")
6: asMethod(object)
5: as(ff, "sparseMatrix")
4: FUN(X[[1L]], ...)
3: lapply(bars, function(x) {
        ff <- eval(substitute(as.factor(fac)[, drop = TRUE], list(fac  
= x[[3]])),
            mf)
        im <- as(ff, "sparseMatrix")
        if (!isTRUE(validObject(im, test = TRUE)))
            stop("invalid conditioning factor in random effect: ",
                format(x[[3]]))
        mm <- model.matrix(eval(substitute(~expr, list(expr = x[[2]]))),
            mf)
        if (rmInt) {
            if (is.na(icol <- match("(Intercept)", colnames(mm))))
                break
            if (ncol(mm) < 2)
                stop("lhs of a random-effects term cannot be an  
intercept only")
            mm <- mm[, -icol, drop = FALSE]
        }
        ans <- list(f = ff, A = do.call(rBind, lapply(seq_len(ncol(mm)),
            function(j) im)), Zt = do.call(rBind,  
lapply(seq_len(ncol(mm)),
            function(j) {
                im at x <- mm[, j]
                im
            })), ST = matrix(0, ncol(mm), ncol(mm), dimnames =  
list(colnames(mm),
            colnames(mm))))
        if (drop) {
            ans$A at x <- rep(0, length(ans$A at x))
            ans$Zt <- drop0(ans$Zt)
        }
        ans
    })
2: lmerFactorList(substitute(foo ~ bar, list(foo = nlform[[2]],
        bar = formula[[3]])), fr, TRUE, TRUE)
1: nlmer(ind.adj ~ SSlogis(trt, Asym, xmid, scal) ~ Asym | clone,
        data = induct, start = c(Asym = 50, xmid = 0.25, scal = 1))


---------------------------------------------------------------------------------
Dr. Andrew Beckerman
Department of Animal and Plant Sciences, University of Sheffield,
Alfred Denny Building, Western Bank, Sheffield S10 2TN, UK
ph +44 (0)114 222 0026; fx +44 (0)114 222 0002
http://www.beckslab.staff.shef.ac.uk/

http://www.flickr.com/photos/apbeckerman/
http://www.warblefly.co.uk