SYSBIONS

SYSBIONS performs Bayesian model selection and parameter inference for systems biology using nested sampling.

Key Features:

• Bayesian Model Comparison: Computes Bayes factors as the ratio of model evidences to quantify relative support for competing models.
• Nested Sampling Methodology: Uses nested sampling to compute model evidence and generate posterior samples in high-dimensional parameter spaces.
• GPU-Accelerated Performance: Implements GPU acceleration to increase computational efficiency for complex biological datasets.
• Prior Sampling with Likelihood Constraints: Provides methods to sample from prior distributions while enforcing likelihood constraints for constrained parameter inference.

Scientific Applications:

• Model selection in systems biology: Identifies the most plausible models among competing hypotheses by quantifying support from empirical data.
• Parameter estimation and posterior analysis: Characterizes posterior distributions of model parameters for downstream analysis and interpretation.
• Predictive modeling: Enables construction of predictive models based on inferred parameter distributions.
• Hypothesis testing and comparison: Facilitates hypothesis testing by comparing model evidences and Bayes factors.

Methodology:

Applies nested sampling to compute model evidences and generate posterior samples, calculates Bayes factors from evidences, supports prior sampling under likelihood constraints, includes optional extensions for handling complex biological datasets, and uses GPU acceleration.