Hi, I'm interested in doing mixed-effects logistic regression on gene expression data (genes in each batch as fixed effects, and different batches as random effects). The data has many times more variables than samples. This calls for a penalised likelihood function, e.g., lasso or ridge. I couldn't find any information about whether lme4 or nlme support penalised regression. Any advice on whether this is possible? Thanks, Gad
Gad Abraham Dept. CSSE and NICTA The University of Melbourne Parkville 3010, Victoria, Australia email: gabraham at csse.unimelb.edu.au web: http://www.csse.unimelb.edu.au/~gabraham