Warning in validityMethod(object) : The following variables have undefined values: rho.1,The following variables have undefined values: rho.2,The following variables have undefined values: rho.3...

[purtyboiii]

I get this error when trying to fit a model using betar or gamma family: Warning in validityMethod(object) : The following variables have undefined values: rho.1,The following variables have undefined values: rho.2,The following variables have undefined values: rho.3...
I'm analysing 6 different genera. The abundance of one of them (Alkanindiges) is the response variable (y) while the abundance of the others are used as explanatory variables (y_all). The y column simply is a subset of the y_all column so for every rows of the explanatory genera, the abundance of Alkanindiges is NA.
Could it be because I have NAs in my response variable? I'm trying to model the influence of the abundance of a few species on the abundance of another species. If it is because of the NAs, how can I ignore them? and could it also be the reason why I can't use an AR(1) function and have to resort to CAR(1)? Thank you for your help!

mod_alk <- mvgam( y ~ ti(y_all, time, by = Genus, bs = 'tp') + s(bioreactor, bs = 're'),
data = APANB_train,
newdata = APANB_test,
family = betar(link = 'logit'),
trend_model = CAR(1),
noncentred = TRUE,
parallel = TRUE,
chains = 4,
backend = 'cmdstanr',
max_treedepth = 14,
adapt_delta = 0.97)

[purtyboiii]

I solved the problem by replacing the NAs with the Alkanindiges's values for the same bioreactor and same day