Dizzy-Beats

Dizzy-Beats performs Bayesian evidence analysis and parameter inference to support model selection and parameter estimation for systems biology and biochemical network models.

Key Features:

• Bayesian Evidence Analysis: Computes Bayesian evidence to quantify support for competing biological models.
• Nested Sampling Algorithm: Implements nested sampling to estimate the logarithm of the Bayesian evidence (Z) and to explore parameter space for complex models.
• Parameter Inference: Calculates moments of model parameters to assess how well data constrain parameter values.
• L1-Norm Likelihood Function: Employs an L1-norm-based likelihood that is applicable to replicated time series data.

Scientific Applications:

• Model Selection: Comparing competing models of biochemical mechanisms in systems biology using Bayesian evidence.
• Parameter Estimation and Uncertainty Quantification: Inferring parameter values and assessing parameter constraint and predictive reliability via parameter moments and parameter distributions.

Methodology:

Uses a Bayesian framework to quantify evidence for models; applies nested sampling to estimate the logarithm of the Bayesian evidence (Z), compute the evidence integral, and explore parameter distributions; and evaluates data likelihoods with an L1-norm function for replicated time series.