I have two treatments and one co-variate (affect the response variable but is also correlated with one of the treatments). Each replicate has been examined for the presence/absence of morphological characters at several time point but there are unfortunately a lot of missing data. What would be the best way of analyzing this data? GLMM? GEE?