Parameters Estimation of Biological Regulatory Networks: An Alternate Approach Based on Computation of Network Centrality

Abstract

Biological systems are complex, diverse and dynamic in nature and these features make them difficult to study. Thus, an easy way is to abstract them in the form of sim ple regulatory networks. Qualitative modeling approaches based on the work of René Thomas are used extensively in the domain of computational systems biology to explore the dynamics of biological regulatory networks. Modeling and analysis on the basis of qualitative modeling framework reveals several behaviors of biological systems in the form of state graphs. These behaviors are driven by certain sets of parameters which are unknown and are very crucial to understand the dynamics of biological systems. There are several approaches which are meant for parameters estimation however, one important problem in these approaches is the exponential number of model parameters. Model checking is one of these approaches based on qualitative modeling framework. It has exponential complexity which when added to complexity of parameters estimation, aggravates the situation in case of large networks; moreover, complex file management and CTL formulas required by model checking approach are difficult to write by people with no programming background, thus, in this work, a simple but scalable approach is proposed to address this challenge by extending the use of betweenness centrality with René Thomas logical formalism to the selection of suitable model parameters. It has lin ear complexity as compared to that of model checking and it is easy to use for everyone (with or without programming background). The developed approach is executed on reported biological regulatory networks for bench marking purpose. This work has been 1 Contents validated by running the approach on a case study of Cerebral Malaria, and comparing its results with those already published in the literature.