Organisms adapt to external perturbations through the optimized structure of their gene regulatory networks (GRNs). In the long-term, the state transition network of a GRN converges to a set of attractors that make the organism resilient to removal or functional impairment of genes. In wireless sensor networks (WSN), such attractors refer to a group of sensors serving as sink nodes for packets sent over multiple hops. This project maps such attractor based genomic robustness onto WSNs to infer optimal topologies and routing strategies that mitigate both sensor failure and a noisy wireless channel. This is being achieved by conducting in silico gene ?knock-down? experiments by simulating the functional removal of a gene from sample GRNs, to understand the dynamics of the attractor state space. This information is next used to design WSN topologies and routing protocols that are resilient to network uncertainty, node breakdown and compromise. This project pursues the design of optimal wiring rules between sensors in a robust WSN that guarantees maximum probability of successful packet transmission under a given routing strategy. The guiding principle is to follow nature?s foot-steps in designing simple rules (i.e., routing algorithms) that guarantee maximum efficiency over an optimized WSN topology. It also develops innovative network-science based tools, and provides insights into the interplay of GRNs and WSNs that inspire new designs for engineered systems (i.e. fault-tolerant topologies for WSNs). Validation and testing are accomplished on real life WSN testbeds. Research results will be disseminated through publications, besides allowing for the design of new graduate-level courses.
Living organisms adapt to external perturbations through the optimized structure of their gene regulatory networks (GRNs). In the long-term, the state transition network of a GRN converges to a set of attractors that make the organisms resilient to removal or functional impairment of genes. In wireless sensor networks (WSN), such attractors refer to a group of sensors serving as sink nodes for packets sent over multiple hops. This project maps such attractor based genomic robustness onto WSNs to infer optimal topologies and routing strategies that mitigate both sensor failure and a noisy wireless channel. This is being achieved by conducting in silico gene knock-down experiments by simulating the functional removal of a gene from sample GRNs, to understand the dynamics of their signal transmission robustness. This information is next used to design logical WSN routing topologies that are resilient to network uncertainty, node breakdown and compromise. The significant results accomplished in this project include: node deployment solutions in bio-inspired WSNs, reverse engineering algorithms to predict the topology of biological networks, and algorithms for predicting network growth based on the design principles of biological networks. Each of these solutions were validated through extensive performance evaluation using simulations as well as a real testbed implementation to demonstrate the high packet transmission robustness and data delivery ratio and low interference and energy consumption of our proposed framework under channel loss and node/link failures. The intellectual merit of the project includes the creation of a robust WSN framework that perform better than most state of the art static routing protocols in WSNs. We have also created an in silico testbed to understand relationships between topology and control of biological networks. The latter contribution is particularly significant as this testbed can now serve to unravel many design principles of biological networks that has been elusive to the experimental biology community. The project also developed innovative network-science based tools, and provided insights into the interplay of GRNs and WSNs that inspire new designs for engineered systems (i.e., fault-tolerant topologies for WSNs). In terms of broader impact, the project findings were incorporated into existing courses taught by the PIs, namely, Advances in Wireless Sensor Networks, Fundamentals of Wireless Networks, and Graph Theory at the University of Texas at Arlington; as well as Parallel Computing course at Virginia Commonwealth University. These courses incorporated several lectures on the proposed research topic and were designed to cater to students from computer science, electrical engineering, bio-engineering, biology and mathematics. Students from these courses were well prepared for networking and biotechnology industry careers alike. The project findings were widely disseminated through several conference and journal publications. Finally, to foster further collaborations and excitement into this project, the PIs have co-founded the International Workshop on Complex Networks Dynamics: Cross-disciplinary tools for modeling, analysis and design (CoNed) organized in 2013 and is expected to be continued.
