MODEL-BASED CLUSTERING WITH DATA CORRECTION FOR REMOVING ARTIFACTS IN GENE EXPRESSION DATA 1