Automated parameter estimation for biological models using Bayesian statistical model checking

摘要: Probabilistic models have gained widespread acceptance in the systems biology community as a useful way to represent complex biological systems. Such are developed using existing knowledge of structure and dynamics system, experimental observations, inferences drawn from statistical analysis empirical data. A key bottleneck building such is that some system variables cannot be measured experimentally. These incorporated into model numerical parameters. Determining values these parameters justify experiments provide reliable predictions when simulations performed research problem. Domain experts usually estimate by fitting Model expressed an optimization problem requires minimizing cost-function which measures notion distance between This often solved combining local global search methods tend perform well for specific application domain. When prior information about available, Bayesian inference commonly used parameter learning. Choosing appropriate technique detailed domain insight underlying system. Using agent-based acute inflammation, we demonstrate novel estimation algorithm discovering amount schedule doses bacterial lipopolysaccharide guarantee set observed clinical outcomes with high probability. We synthesized twenty-eight unknown parameterized instantiated satisfies four specifications describing dynamic behavior model. new algorithmic stochastic given behavioral written formal mathematical logic. Our uses checking, sequential hypothesis testing, automatically synthesize probabilistic models.