LooplessFluxSampler

LooplessFluxSampler explores the loopless mass-balanced flux solution space of metabolic models using an Adaptive Directions Sampling on a Box (ADSB) algorithm.

Key Features:

• ADSB Algorithm: Implements Adaptive Directions Sampling on a Box, rooted in the general Adaptive Direction Sampling framework with Parallel ADS theoretical convergence and irreducibility results.
• Efficient Space Traversal: Samples directions that adapt to the target distribution, achieving faster mixing than other methods by traversing the sample space more efficiently.
• Uniform Distribution Guarantee: Targets the uniform distribution over convex regions and provably converges on this distribution over more general non-convex regions provided the sample can have full support.
• Markov Chain Diagnostics Suite: Includes diagnostics for assessing the quality of the final sample and the performance of the algorithm.

Scientific Applications:

• Metabolic Network Analysis: Provides unbiased appraisal of the capabilities of metabolic networks through uniform random sampling of mass-balanced flux solutions.
• Statistical Inference: Enables scalable statistical inference of the loopless mass-balanced solution space of large metabolic models.
• Loopless Flux Space Exploration: Explores the properties of the almost surely non-convex loopless flux space without thermodynamically infeasible loops.

Methodology:

The tool uses Adaptive Directions Sampling on a Box (ADSB), rooted in the general Adaptive Direction Sampling (ADS) framework, specifically Parallel ADS. ADSB samples directions that adapt to the target distribution to traverse the sample space. The algorithm targets the uniform distribution over convex regions and converges on this distribution over non-convex regions with full support.