Ahmed Abdelzaher

Performance of wireless sensor topologies inspired by E. coli genetic networks

Wireless Sensor Networks (WSNs) form a critical component in modern computing applications; given their size, ability to process and communicate information, and to sense stimuli, they are a promising part of The Internet of Things. However, they are also plagued by reliability and node failure problems. Here we address these problems by using E. coli Gene Regulatory Networks (GRNs) - believed to be robust against signaling disruptions, such as gene failures - to study the transmission properties of randomly-generated WSNs and transmission structures derived from these genetic networks. Selection of sink nodes is crucial to the performance of these networks; here we introduce two sink-node selection techniques: a Motif-based, and a Highest Degree-based approach. Using NS-2 simulations, the performance of both networks is evaluated under varying channel loss models. Successful packet receipts are compared among these networks, which are shown to be higher using GRNs for the communication structure, rather than randomly generated WSNs. This work paves the way for future development of fault-tolerant and robust WSN deployment and routing algorithms.

Leveraging the robustness of genetic networks: a case study on bio-inspired wireless sensor network topologies

Wireless sensor networks (WSNs) form a critical component in modern computing applications; given their size, ability to process and communicate information, and to sense stimuli, they are a promising part of the Internet of Things. However, they are also plagued by reliability and node failure problems. Here we address these problems by using the Gene Regulatory Networks (GRNs) of the organism Escherichia coli—believed to be robust against signaling disruptions, such as gene failures—to study the transmission properties of randomly-generated WSNs and transmission structures derived from these genetic networks. Selection of sink nodes is crucial to the performance of these networks; here we introduce four sink-node selection techniques: two motif-based, an attractor based and a highest degree-based approach and perform comprehensive simulations to assess their performance. Specifically, we use NS-2 simulations to evaluate the packet transmission robustness properties of such GRN-derived communication structures as against typical randomly deployed sensor network topologies under varying channel loss models. Packet receipt rates are compared among these networks, which are shown to be higher using GRNs for the communication structure, rather than randomly generated WSNs. We also evaluate the performance of communication structures derived from existing biological network generation models to assess their applicability in providing robust communication. This work paves the way for future development of fault-tolerant and robust WSN deployment and routing algorithms based on the inherent signal transmission robustness properties of the gene regulatory network topologies.

Motif Participation by Genes in E. coli Transcriptional Networks.

Motifs are patterns of recurring connections among the genes of genetic networks that occur more frequently than would be expected from randomized networks with the same degree sequence. Although the abundance of certain three-node motifs, such as the feed-forward loop, is positively correlated with a networks’ ability to tolerate moderate disruptions to gene expression, little is known regarding the connectivity of individual genes participating in multiple motifs. Using the transcriptional network of the bacterium Escherichia coli, we investigate this feature by reconstructing the distribution of genes participating in feed-forward loop motifs from its largest connected network component. We contrast these motif participation distributions with those obtained from model networks built using the preferential attachment mechanism employed by many biological and man-made networks. We report that, although some of these model networks support a motif participation distribution that appears qualitatively similar to that obtained from the bacterium E. coli, the probability for a node to support a feed-forward loop motif may instead be strongly influenced by only a few master transcriptional regulators within the network. From these analyses we conclude that such master regulators may be a crucial ingredient to describe coupling among feed-forward loop motifs in transcriptional regulatory networks.

PANNOTATOR: an automated tool for annotation of pan-genomes.

Due to next-generation sequence technologies, sequencing of bacterial genomes is no longer one of the main bottlenecks in bacterial research and the number of new genomes deposited in public databases continues to increase at an accelerating rate. Among these new genomes, several belong to the same species and were generated for pan-genomic studies. A pan-genomic study allows investigation of strain phenotypic differences based on genotypic differences. Along with a need for good assembly quality, it is also fundamental to guarantee good functional genome annotation of the different strains. In order to ensure quality and standards for functional genome annotation among different strains, we developed and made available PANNOTATOR (http://bnet.egr.vcu.edu/iioab/agenote.php), a web-based automated pipeline for the annotation of closely related and well-suited genomes for pan-genome studies, aiming at reducing the manual work to generate reports and corrections of various genome strains. PANNOTATOR achieved 98 and 76% of correctness for gene name and function, respectively, as result of an annotation transfer, with a similarity cut-off of 70%, compared with a gold standard annotation for the same species. These results surpassed the RAST and BASys softwares by 41 and 21% and 66 and 17% for gene name and function annotation, respectively, when there were reliable genome annotations of closely related species. PANNOTATOR provides fast and reliable pan-genome annotation; thereby allowing us to maintain the research focus on the main genotype differences between strains.

Feature ranking in transcriptional networks: packet receipt as a dynamical metric

Machine learning techniques may be useful in determining the features contributing to some biological properties, such as robustness, which is the tendency for biological systems to resist a change of state. In this work, we compare transcriptional subnetworks extracted from the bacterium Escherichia coli and the bakers yeast Saccharomyces cerevisiae using in silico experiments. We use the packet receipt rate as a metric to quantify biological robustness, which is different from the usual structural metrics since it captures the dynamic behavior of the network. We define seventeen features based on structural significance, such as transcriptional motifs, and conventional metrics, such as average shortest path and network density, among others. Feature ranking is performed, based on a grid-search method to identify Support Vector Machine classifier parameters using cross validation. Our results indicate that feed-forward loop based features are important for bacterial transcriptional networks, whereas network density, degree-centrality based and bifan-based features are found to be significant for yeast-derived transcriptional networks. Interestingly, results suggest that feature significance varies with network size (number of nodes). As a first, this study quantifies the impact of the feed-forward loop and bifan transcriptional motif abundance observed in natural networks.

Crosstalk and the Dynamical Modularity of Feed-Forward Loops in Transcriptional Regulatory Networks

Network motifs, such as the feed-forward loop (FFL), introduce a range of complex behaviors to transcriptional regulatory networks, yet such properties are typically determined from their isolated study. We characterize the effects of crosstalk on FFL dynamics by modeling the cross regulation between two different FFLs and evaluate the extent to which these patterns occur in vivo. Analytical modeling suggests that crosstalk should overwhelmingly affect individual protein-expression dynamics. Counter to this expectation we find that entire FFLs are more likely than expected to resist the effects of crosstalk (≈20% for one crosstalk interaction) and remain dynamically modular. The likelihood that cross-linked FFLs are dynamically correlated increases monotonically with additional crosstalk, but is independent of the specific regulation type or connectivity of the interactions. Just one additional regulatory interaction is sufficient to drive the FFL dynamics to a statistically different state. Despite the potential for modularity between sparsely connected network motifs, Escherichia coli (E. coli) appears to favor crosstalk wherein at least one of the cross-linked FFLs remains modular. A gene ontology analysis reveals that stress response processes are significantly overrepresented in the cross-linked motifs found within E. coli. Although the daunting complexity of biological networks affects the dynamical properties of individual network motifs, some resist and remain modular, seemingly insulated from extrinsic perturbations—an intriguing possibility for nature to consistently and reliably provide certain network functionalities wherever the need arise.

Quantifying robustness in biological networks using NS-2

Biological networks are known to be robust despite signal disruptions such as gene failures and perturbations. Extensive research is currently under way to explore biological networks and identify the underlying principles of their robustness. Structural properties such as power-law degree distribution and motif abundance have been attributed for robust performance of biological networks. Yet, little has been done so far to quantify such biological robustness. We propose a platform to quantify biological robustness using network simulator (NS-2) by careful mapping of biological properties at the gene level to that of wireless sensor networks derived using the topology of gene regulatory networks found in different organisms. A Support Vector Machine (SVM) learning model is used to measure the correlation of packet transmission rates in such sensor networks. These sensor networks contain important topological features of the underlying biological network, such as motif abundance, node/gene coverage, and transcription-factor network density, which we use to map the SVM features. Finally, a case study is presented to evaluate the NS-2 performance of two gene regulatory networks, obtained from the bacterium Escherichia coli and the baker’s yeast Sachharomyces cerevisiae.

Contribution of canonical feed-forward loop motifs on the fault-tolerance and information transport efficiency of transcriptional regulatory networks

Abstract Motifs and degree distribution in transcriptional regulatory networks play an important role towards their fault-tolerance and efficient information transport. In this paper, we designed an innovative in silico canonical feed-forward loop motif knockout experiment in the transcriptional regulatory network of E. coli to assess their impact on the following five topological features: average shortest path, diameter, closeness centrality, global and local clustering coefficients. Additional experiments were conducted to assess the effects of such motif abundance on E. coli ’s resilience to nodal failures and the end-to-end transmission delay. The purpose of this study is two-fold: (i) motivate the design of more accurate transcriptional network growing algorithms that can produce similar degree and motif distributions as observed in real biological networks and (ii) design more efficient bio-inspired wireless sensor network topologies that can inherit the robust information transport properties of biological networks. Specifically, we observed that canonical feed forward loops demonstrate a strong negative correlation with the average shortest path, diameter and closeness centralities while they show a strong positive correlation with the average local clustering coefficient. Moreover, we observed that such motifs seem to be evenly distributed in the transcriptional regulatory network; however, the direct edges of multiple such motifs seem to be stitched together to facilitate shortest path based routing in such networks.

Transcriptional Network Growing Models Using Motif-Based Preferential Attachment

Understanding relationships between architectural properties of gene-regulatory networks (GRNs) has been one of the major goals in systems biology and bioinformatics, as it can provide insights into, e.g., disease dynamics and drug development. Such GRNs are characterized by their scale-free degree distributions and existence of network motifs – i.e., small-node subgraphs that occur more abundantly in GRNs than expected from chance alone. Because these transcriptional modules represent “building blocks” of complex networks and exhibit a wide range of functional and dynamical properties, they may contribute to the remarkable robustness and dynamical stability associated with the whole of GRNs. Here, we developed network-construction models to better understand this relationship, which produce randomized GRNs by using transcriptional motifs as the fundamental growth unit in contrast to other methods that construct similar networks on a node-by-node basis. Because this model produces networks with a prescribed lower bound on the number of choice transcriptional motifs (e.g., downlinks, feed-forward loops), its fidelity to the motif distributions observed in model organisms represents an improvement over existing methods, which we validated by contrasting their resultant motif and degree distributions against existing network-growth models and data from the model organism of the bacterium Escherichia coli. These models may therefore serve as novel testbeds for further elucidating relationships between the topology of transcriptional motifs and network-wide dynamical properties.

Mixed Degree-Degree Correlations in Directed Social Networks

Many complex networks exhibit homophilic, or assortative degree mixing–the tendency for networked nodes to connect with others of similar degree. For social networks, this phenomenon is often referred to colloquially by the mantra ‘your friends have more friends than you do.’ We analyzed datasets for 16 directed social networks, and report that some of them exhibit both assortative (positive correlations) and disassortative (negative correlations) degree mixing across the totality of their degrees. We show that this mixed trend can be predicted based on the value of Pearson correlations computed for the directed networks. This stands in contrast to previous results reported for social networks that mark them as purely assortative. Finally, we discuss mechanisms by which these trends emerge from random models of network creation.