An intuitive method to visualize Boolean network attractors landscape based on probability distribution from trajectories

Abstract

MotivationBoolean network modeling is a coarse-grained modeling approach that has been highly valued for its simple abstract representation of biological regulation in molecular network. The simplified On and Off expression of gene or protein product is a qualitative representation of experimental findings for gene gets turned on or turned off. Boolean network constructed can be analyzed as having point attractor or cyclic attractor; however, current softwares can visualize the Boolean network attractors on a plane and not visualizing Boolean attractors landscape in a 3 dimensional view. ResultsHere, we propose an intuitive method to visualize Boolean network attractors land-scape based on probability distribution P of all possible trajectories in the state transition graph. We formulate a quasi-potential value U = - ln P that enables us to plot the Boolean network attractors landscape in 3D. There are two types of attractors in Boolean network: cyclic attractor and point attractor. Our intuitive method can visualize Boolean network in a landscape with point attractor (node labeled with yellow color) and cyclic attractor (node labeled with pink color). In the 3D landscape the attractors are located at the bottom of the directed graphs. The propose method can be applied in the study of networks in general, for example applications in directed graph, social network for popular users or e-commerce store to find favorite products as identified by attractors. Availability and implementationThe source code of Python implementation of the propose method can be downloaded from Boolean-net-attractor-Land URL link.