Hi,

I look for the best model fitting with my data. I have measured different genetic diversity indices (Ar, He, number of breeders, Fis) of a zoonotic pathogen . Every 4 are continuous, two of them are normals and 2 are not normals. These dependant variables are correlated. I wan't to know if these variables could be explained by 4 independant variables : Host sex (categorical : male/female - only for human), Site of transmission (categorical: site1-2-3), Host species (categorical : Human/Animal), host age (continuous - only for human).

What could be the best model ?

Multivariate GLM with the 4 dependant variables simultaneously ? (problem some variables are not normally distributed)

Generalized Linear Model for each dependant variable taken independently ?

....Others ?

Every comments, suggestion are welcome !

R command or SPSS procedure is also welcome

Thank you