SEEDS

SEEDS infers structural model errors and estimates unknown inputs in ordinary differential equation (ODE) dynamic models to improve mechanistic model accuracy in systems biology.

Key Features:

• Dynamic Model Analysis: Operates on dynamic models represented as ordinary differential equations (ODEs) to capture mechanistic and causal interactions in biological systems.
• Inference of Structural Errors: Identifies discrepancies in model structure and functional form that reduce predictive accuracy.
• Estimation of Unknown Inputs: Estimates hidden or unknown environmental inputs that act as controls to minimize differences between predicted and observed outputs.
• Algorithmic Approach: Implements two recently developed algorithms for inferring both structural errors and unknown inputs, rooted in the Dynamic Elastic Net concept.
• Model Recalibration and Experimental Design Support: Provides information to guide recalibration of model parameters and to inform the design of subsequent experiments.
• Implementation and Interoperability: Provided as an R package with compatibility for SBML (Systems Biology Markup Language) via the rsbml package.

Scientific Applications:

• Pharmacology: Improves mechanistic model accuracy to support drug response prediction and intervention strategies.
• Genetics: Refines dynamic genetic network models by correcting structural errors and accounting for unknown inputs.
• Cellular Biology: Enhances cellular pathway and signaling models for more reliable mechanistic interpretation.
• Prediction and Experimental Optimization: Enables more reliable predictions and interventions, supporting better-targeted therapies, optimized experimental protocols, and deeper mechanistic understanding.

Methodology:

Integrates output measurements with ODE-based dynamic models using two algorithms based on the Dynamic Elastic Net to infer structural errors and estimate unknown inputs by iterative adjustment of model parameters, and supports SBML model import via the rsbml R package.