# lme when model doesn't follow a normal distrubution (nlme library)

I am trying to do lme with the library nlme as it is good for ML. If I define the model as:

model<- lme(GHCB~status + age + years_since, random= ~1|Patient/ID, data = cube, na.action = na.omit)

qqnorm(resid(model))

qqline(resid(model))

If the model doesn't fit normality, and after assessing all the variables to see which one doesn't have a normal distribution with

shapiro.test(database\$GHCB)

I see that none of them is normally distributed. So how should I proceed?

I've tried to log or ^2 the variables but I am not sure how to convert the distribution in a normal distribution, hence, the model is not working.

What do I need to do now? try to convert the variables in "normal" or change the model? and if so, how should I do that?

Thank you very much!! Lili

PS: I attach the qqnorm for the model

• see robustlmm package for R ... ? also, check for interactions – Ben Bolker Jun 23 '18 at 2:01