Related Experiment Videos
DISTING: A web application for fast algorithmic computation of alternative indistinguishable linear compartmental
Natalie R Davidson1, Keith R Godfrey2, Faisal Alquaddoomi1
1Biocybernetics Laboratory, Computer Science Department, University of California Los Angeles, CA 90095, U.S.A.
Computer Methods and Programs in Biomedicine
|April 11, 2017
Summary
DISTING is a new web application that identifies alternative, input-output indistinguishable compartmental models. It simplifies complex analysis, making it easier to find structurally identifiable models.
Area of Science:
- Pharmacokinetics and Pharmacodynamics
- Systems Biology
- Mathematical Modeling
Background:
- Assessing structural identifiability of linear compartmental models is crucial for reliable analysis.
- Existing computational tools primarily focus on identifiability, not indistinguishability.
- Input-output indistinguishability poses a challenge in model selection and interpretation.
Purpose of the Study:
- Introduce DISTING, a novel web application for computing structurally identifiable linear compartmental models that are input-output indistinguishable.
- Provide a user-friendly interface for assessing model indistinguishability, a capability not previously available in computer packages.
- Illustrate the application of DISTING using examples of varying complexity.
Main Methods:
- Utilizes advanced geometric and algebraic properties of linear compartmental models.
- Features a user-friendly graphical model interface.
- Incorporates novel computational tools for accelerated analysis, including Jacobian matrix reduction and submatrix rank reduction.
Main Results:
- DISTING successfully identified multiple input-output indistinguishable models for 2, 3, and 4-compartment systems.
- A 2-compartment model yielded two indistinguishable models.
- A 4-compartment model revealed five indistinguishable models, highlighting the impact of unobserved compartments.
Conclusions:
- DISTING is freely available online, offering a significant advancement in compartmental modeling analysis.
- The application simplifies complex algebraic analysis, previously performed manually.
- Facilitates the identification of alternative, indistinguishable models, improving model selection and understanding.