PyDREAM

PyDREAM performs high-dimensional parameter inference and uncertainty estimation for biological models by implementing the (Multiple-Try) Differential Evolution Adaptive Metropolis (DREAM(ZS)) algorithm.

Key Features:

• Algorithm: Implements the (Multiple-Try) Differential Evolution Adaptive Metropolis (DREAM(ZS)) algorithm for MCMC sampling.
• Parameter inference: Calibrates model parameters against experimental data to estimate hard-to-measure parameters.
• Multivariate posterior estimation: Performs multivariate posterior model parameter distribution estimation.
• Uncertainty estimation: Quantifies parameter uncertainty from posterior distributions.
• Distributed computing: Leverages distributed computing resources to scale inference for CPU-intensive models.
• High-dimensional robustness: Targets robust performance in complex, parameter-rich models and mitigates slow or premature convergence typical of traditional MCMC in high-dimensional spaces.
• Implementation: Implemented in Python.

Scientific Applications:

• Systems biology: Calibration and uncertainty analysis of mechanistic systems biology models.
• Ecological modeling: Parameter inference and uncertainty quantification for ecological models.
• Complex model calibration: Parameter inference and uncertainty estimation for other fields requiring calibration of complex biological models.

Methodology:

Implements the (Multiple-Try) Differential Evolution Adaptive Metropolis (DREAM(ZS)) algorithm; performs multivariate posterior model parameter distribution estimation; calibrates parameters against experimental data; leverages distributed computing for scalable inference and addresses convergence issues of traditional MCMC in high-dimensional parameter spaces to enable uncertainty estimation.