Joint model fitting to data on separate enantiomers and enantiomeric mixtures

All, We have a data set comprising three measurements per time point, namely the separate R- and S-concentrations and the total concentration (R+S) of a chiral drug given as the racemic mixture. We would like to fit a model to all data simultaneously. The plan is to estimate one set of parameters for the S- enantiomer and one set for the R- enantiomer, using the sum of S- and R concentrations to predict R+S; defining CMTs 1, 3, and 5 as observational compartments for R-, S-, and RS, respectively, we would like to model A(5)=A(1)+A(3). Any suggestion on how to do proceed with this in ADVAN5 or ADVAN6 would be appreciated! / Johan and Anna