Parameter estimation in large-scale systems biology models: a parallel and self-adaptive cooperative strategy.

Area of Science:

• Computational systems biology
• Bioinformatics
• Computational modeling

Background:

• Parameter estimation in nonlinear dynamic models is crucial for computational systems biology.
• Global optimization methods are computationally expensive and require parameter tuning.
• Large-scale kinetic models are essential for understanding biological systems.

Purpose of the Study:

• To present a novel parallel method, self-adaptive cooperative enhanced scatter search (saCeSS), for accelerating parameter estimation.
• To address the computational cost and tuning challenges of existing optimization methods.
• To facilitate the development of large-scale dynamic models.

Main Methods:

• Developed saCeSS, a parallel method based on scatter search metaheuristic.
• Incorporated asynchronous cooperation between parallel processes.
• Implemented coarse and fine-grained parallelism with self-tuning strategies.

Main Results:

• saCeSS demonstrated robust and efficient performance on diverse kinetic models (E. coli, yeast, fly, CHO cells).
• Achieved significant reductions in computation time (days to minutes) compared to state-of-the-art methods.
• Showcased efficiency even with a small number of processors.

Conclusions:

• saCeSS enables solving medium to large-scale parameter estimation problems efficiently.
• The method's self-tuning mechanisms simplify its use for non-experts.
• saCeSS is poised to play a key role in advancing large-scale and whole-cell dynamic modeling.