Sleep’s quality — t0psqi • mglm4twin

Data set concerning sleep’s quality in a sample of 250 (135 DZ and 116 MZ) Danish twin pairs. The traits are cortisone levels when waking up (T0) and PSQI (Pittsburgh Sleep Quality Index). It is a simulated data set based on the parameter estimates obtained fitting the model to a motivating real data set. The code for the simulation is available in the folder data-raw.

Twin_pair - Code of twin within the pair (1 and 2).
Twin_id - Twin code.
Age - Twin age.
Type - Twin zygosity (DZ - dizygotic; MZ - monozygotic).
Gender - Gender (Male and Female).
Group - Treatment group, it is categorical covariate for composing the linear predictor.
T0 - Cortisone levels when waking up (continuous trait).
PSQI - Pittsburgh Sleep Quality Index (bounded trait) divided by 21 (scale maximum).

Usage

data(t0psqi)

Format

a data.frame with 502 records and 8 variables.

Source

Bonat, W. H. and Hjelmborg, J. v. B. (2020) Multivariate Generalized Linear Models for Twin and Family data. to appear.

Examples

require(mglm4twin)
form_T0 <- T0 ~ Age + Gender + Group + Type*Twin_pair
form_PSQI <- PSQI ~ Age + Gender + Group + Type*Twin_pair
AE <- mt_twin(N_DZ = 135, N_MZ = 116, n_resp = 2, model = "AE")
fit_AE <- mglm4twin(linear_pred = c(form_T0,  form_PSQI),
                   matrix_pred = AE,
                   link = c("log","logit"),
                   variance = c("tweedie","binomialP"),
                   control_algorithm = list(tuning = 0.25, max_iter = 100),
                   power_fixed = c(FALSE,FALSE), data = t0psqi)
#> Warning: non-integer counts in a binomial glm!
#> Automatic initial values selected.