Background and Resources

(1) Mamcat Estimation Simulation
(2) Contributors
(3) Contacts

W3MAMCAT Parameter Estimator

1. MODELING: A linear n-compartment MAMMILLARY and/or CATENARY model can be formulated and fully quantified by tranformation from a known MULTIEXPONENTIAL MODEL (sum of n exponentials). Solutions are limited to models with both input and output in compartment 1, and n must be no greater than 10.

NOTE: A 2-step modeling process is assumed. The multiexponential model is typically obtained from kinetic time-series data using a data fitting program like DIMSUM+, another expert system available from the Biocybernetics Laboratory.

2. MODEL DISCRIMINATION/DISTINGUISHABILITY: An expert subsystem is included for determining whether a mammillary can be distiguished form a catenary linear compartmental model, each with n compartments, but input and output may be in any (same of different) compartments. We assume both can be fitted to a single output data set; the questions is whether the two different structures can be distinguished bia this data. The theory is fully developed pedagogically, along with specific solutions.

Contributors

Acknowledgements

The author would like to thank the following contributors for their thoughtful suggestions for improvement and their careful examination of the functionality and utility of this W3MAMCAT Parameter Estimator simulation package:

Principal Investigator: Joseph J. DiStefano III, Ph.D.
Professor of Computer Science and Medicine
University of California, Los Angeles
Professor DiStefano suggested this web based version of his biocybernetics lab's MAMCAT+ as a M.S. project. During the whole process, Prof. DiStefano has been an invaluable source of knowledge as to the functionality and usability of W3MAMCAT. His steadfastness for producing the most easily usable application, while always keeping in mind the potential student user, has continually improved this application since its conception.

Contributor: Kirk Bulis, MS
University of California, Los Angeles
This project was based in large part on the work of Kirk Bulis, and his Monte Carlo Simulator (W3MCSim). A huge amount of thanks must be given to Kirk, for his guidance, vast knowledge and willingness to help at every point in design of W3MAMCAT. Without his tireless help and dedication, the completion this project would not of been possible.

Resources

Box, Don, Essential COM, Addison Wesley Longman, Inc.,
Reading, MA, 1998.
Chappell, David, Understanding ActiveX™ and OLE, Microsoft
Press, Redmond, WA, 1996.
J.J. DiStefano, III, Complete Parameter Bounds Quasiidentifiability Conditions for a Class of Unidentifiable Linear Systems, Mathematical Biosciences 65:51-68 (1983)
E.M. Landaw, B. C. Chen, J. J. DiStefano, III, An Algorithm for the Identifiable Parameter Combinations of the General, Mathematical Biosciences. 72:199-212 (1884)
B. C. Chen, E. M. Landaw, J.J. DiStefano, III, Algorithms for the Identifiable Parameter Combinations and Parameter Bounds of Unidentifiable Catenary Compartmental Models, Mathematical Biosciences. 76:59-68 (1985)
J.J. DiStefano III, B. C. Chen, E.M. Landaw, Pool Size and Mass Flux Bounds and Quasiidentifiability Relations for Catenary Models, Mathematical Biosciences, 88:1-14 (1988)
R. Lindell, J. J. DiStefano, E. M. Landaw, Statistical Variability of Parameter bounds for n-Pool Unidentifiable Mammillary and Catenary Compartmental Models. Mathematical Biosciences 91:175-199 (1988)
A. B. Kuhn de Chizelle, J.J. DiStefano III, MAMCAT: An Expert System for Distinguishing Between Mammillary and Catenary Compartmental MOdels. Comput. Biol. Mel. 24:189-204 (1994)
Paolo Vicini, Hsiao-Te Su, J. J. DiStefano III, Interval Identifiability of Mammillary and Catenary Compartmental Models with Some Known Rate Constants. (Submitted)

Previous Work

W3MAMCAT is the online version of MAMCAT+ Version 1.0, principally programmed by Hsiau-Te Su. The mathematical computations are done using the same code as the original program. The expert system for distinguishing mammillary and catenary models was originally programmed by Anh-Ngoc B. Kuhn De Chizelle. The expert system was translated into JavaScript for inclusion in W3MAMCAT.

Contacts

Joseph J. DiStefano III
Professor of Computer Science, Medicine, and Biomedical Engineering
UCLA
4711 Boelter Hall
Los Angeles, CA 90095-1596
telephone: (310) 825-7482
fax: (310) 825-2273
email: joed@cs.ucla.edu

Solomon Russell
Graduate Student
UCLA, Computer Science Department
email: solomon@cs.ucla.edu

Chance Risner
Student
El Camino College, Computer Science Department
email: chance_risner@elcamino.edu

UCLA Biocybernetics Laboratory
web: http://www.biocyb.cs.ucla.edu