Repeated measurements using relmatGlmer

[MelLinde]

Hi,

I am trying to analyse count data with relmatGlmer accounting for relatedness and repeated measurements. When including repeated measurements I get the error “nrow(Ztlist[[i]])%%ncol(Ztlist[[i]]) == 0 is not TRUE”.

When I transform the example data “dat40” into count data (trait1, trait2) relmatGlmer works (but outputs warnings). When I then add repeated measurements to the example data, relmatGlmer doesn’t work and outputs the same error as mentioned above.

Here is the code I run:

##########
library(lme4) # needed for VarCorr function
library(lme4qtl)

data(dat40)

data.count <- read.table("data.count.txt", sep="\t")
data.count.repeat <- read.table("data.count.repeat.txt", sep="\t")

No repeated measurements (works):

relmatGlmer(cbind(trait1, trait2) ~ AGE + SEX + (1|FAMID) + (1|ID), data.count, relmat = list(ID = kin2), family = binomial)

relmatGlmer((trait1/(trait1+trait2)) ~ AGE + SEX + (1|FAMID) + (1|ID), data.count, weights=(trait1+trait2), relmat = list(ID = kin2), family = binomial)

Repeated measurements (does not work):

mod <- relmatGlmer(cbind(trait1, trait2) ~ AGE + SEX + (1|FAMID) + (1|ID), data.count.repeat, relmat = list(ID = kin2), family = binomial)

mod <- relmatGlmer((trait1/(trait1+trait2)) ~ AGE + SEX + (1|FAMID) + (1|ID), data.count.repeat, weights=(trait1+trait2), relmat = list(ID = kin2), family = binomial)

##########

Please find the input data attached. Any hint how to model data with relamtGlmer including repeated measurements would help me a lot!

Many thanks and best regards,
Melanie

data.count.repeat.txt
data.count.txt

[variani]

Hi Melanie,

I started to address this issue using data from the pedigreemm R package, which also have repeated measurements.

Now I remember that I faced this problem in the vignette by running lmer and fixed it that time. It assumed to be straightforward to do the same fix for glmer. Let me a bit more time.

Best,
Andrey

[MelLinde]

Hi Andrey,

many thanks for the update. But pleas remember that I also need to include a relatedness/kinship matrix. So, need to work with relmatGlmer.

Best,
Melanie

[variani]

Hi Melanie,

It should now work. I put some relevant R scripts here https://github.com/variani/lme4qtl/tree/master/misc/issues/05.

Best,
Andrey

[MelLinde]

Hi Andrey,

yes, it works fine. Thanks a lot!

Best,
Melanie