Modeling 2 vs 3 compartments for mutant proteins

I've been asked to model zymogen and activated wild type and mutated forms of an enzyme in in-bred mice. So one can consider the mutation of the administered protein a covariate, if you will For wild-type protein the model is 3 compartment, but when I throw in various other mutations (8 in all) the model changes from 3- to 2- compartments, depending upon the mutated protein administered. Normally we decide what number of compartments (exponents) explains our model, then look at etas and covariate. Has anyone set up a model where the base-model fit is 2- vs 3- compartments depending upon what "drug" (or crystalline form or isotopic formulation) was administered? I am hoping that the model for each mutation may be partially explained by its binding (or lack thereof) to circulating activating or quenching proteins. Thanks Paul Paul R. Hutson, PharmD, BCOP Distinguished Professor (CHS) Thora M. Vervoren Professor for Research in Psychoactive Substances UW School of Pharmacy Director, UW Madison Transdisciplinary Center for Research in Psychoactive https://www.research.pharmacy.wisc.edu/tcrps Faculty Leader, Paul P. Carbone Comprehensive Cancer Center Cancer Pharmacology https://cancer.wisc.edu/research/resources/ddc/cancer-pharmacology/ T: 608.263.2496 [email protected]