BioPreDyn-Bench

BioPreDyn-Bench provides a suite of benchmark parameter estimation and optimal experimental design problems for dynamic modeling of medium- to large-scale kinetic models in systems biology.

Key Features:

• Comprehensive Problem Set: Contains six parameter estimation problems derived from medium- to large-scale kinetic models representing metabolism, transcription, signal transduction, and development.
• Multiple Implementations: Each benchmark problem is provided in several implementations and formats to support use across different computational environments.
• Pre-computed Results: Includes pre-computed results obtained using specific solvers to enable validation and comparison of methods.
• Analytical Support: Supplies basic analysis and interpretation material accompanying each problem for assessing model performance.
• Versatile Applications: Applicable to parameter estimation, sensitivity and identifiability analyses, model reduction techniques, and optimal experimental design benchmarking.

Scientific Applications:

• Escherichia coli models: Provides a benchmark problem derived from Escherichia coli for metabolic dynamic modeling and parameter estimation.
• Saccharomyces cerevisiae models: Provides a benchmark problem derived from Saccharomyces cerevisiae for dynamic modeling and parameter inference.
• Drosophila melanogaster models: Provides a developmental modeling benchmark derived from Drosophila melanogaster.
• Chinese Hamster Ovary cell models: Contains a benchmark representing Chinese Hamster Ovary cells for dynamic kinetic modeling.
• Generic signal transduction network: Includes a generic signal transduction network model for benchmarking signaling pathway dynamics.
• Methodological benchmarking: Supports systematic evaluation and comparison of computational methods and software for dynamic modeling.

Methodology:

Formulates medium- to large-scale kinetic dynamic models and provides benchmark parameter estimation and optimal experimental design problems with pre-computed solver results and support for sensitivity and identifiability analyses and model reduction techniques.