A distributed software package for global sensitivity analysis of biological models

Mathematical modeling is a vital tool for studying biological systems. Due to the system complexity and technical challenges in molecular level measurement, it is commonly the case that a large number of model parameters have uncertain values. Analyzing model dynamics from a single estimated parameter set is insufficient and liable for misleading results. Global sensitivity analysis (GSA) has been recommended as a must-have step in the process of developing reliable models. However, the technique comes at high computation costs as it is based on Monte Carlo simulation which requires a large number of model evaluations and manipulating on massive data for sensitivity estimation. In this work, we develop a software package for global sensitivity estimation of biological models. The software is deployed on KISTI high performance computing (HPC) cluster environment to provide web-service to system biology modelers.