The use and analysis of multiple responses in multicompartment cellular systems

Within multicompartment cellular systems in the steady state the distrubution of tracer as a function of time is described by sums of exponentials. A design for data collection and analysis is set forward to deal with the problems of lumping and ill-conditioning inherent in such a mathematical description. This design requires the provision of multiple responses and can be analyzed by a Bayesian multivariate technique. Although only one response may be capable of completely characterizing the system, there is much to be gained from additional responses in terms of precision in parameter estimation. A time series is used to account for autocorrelated error in the responses. The design is compared to other least squares approaches and is demonstrated in an investigation of the transport and storage of serotonin in blood platelets. The kinetic parameters describing serotinin movement are estimated and their joint confidnece regions plotted.