Bulletin of the American Physical Society
APS March Meeting 2023
Volume 68, Number 3
Las Vegas, Nevada (March 5-10)
Virtual (March 20-22); Time Zone: Pacific Time
Session Q02: Statistical Physics of Networks: Theory and ApplicationsFocus
|
Hide Abstracts |
Sponsoring Units: GSNP DBIO Chair: Guido Caldarelli, University of Venice Ca'Foscari Room: Room 125 |
Wednesday, March 8, 2023 3:00PM - 3:36PM |
Q02.00001: Influence maximization in Boolean networks Invited Speaker: Filippo Radicchi The optimization problem aiming at the identification of minimal sets of nodes able to drive the dynamics of Boolean networks toward desired long-term behaviors is central for some applications, as for example the detection of key therapeutic targets to control pathways in models of biological signaling and regulatory networks. Unfortunately, the complexity of the optimization problem is exponential, making it exactly solvable on very small systems only. Some scalable approaches exist, but they rely on linear approximations; other approaches estimate nonlinear effects, but they are generally not scalable. In this talk, I will introduce an alternative method inspired by those used in the solution of the well-studied problem of influence maximization for spreading processes in social networks. The computational time of the proposed method scales cubically with the network size. This is achieved thanks to some strong approximations, as for example neglecting dynamical correlations among Boolean variables. However, the method has the desirable feature of fully accounting for the nonlinear nature of Boolean dynamics. I will validate the method on small gene regulatory networks whose dynamical landscapes are known by means of brute-force analysis. I will then systematically apply it to a large collection of gene regulatory networks revealing that for about 65% of the analyzed networks, the minimal driver sets contain less than 20% of their nodes. |
Wednesday, March 8, 2023 3:36PM - 3:48PM |
Q02.00002: Percolation in Borromean Networks Donald Ferschweiler, Ryan Blair, Alexander R Klotz Inspired by experiments on topologically linked DNA networks, we consider the connectivity of Borromean networks, in which no two rings share a pairwise-link, but groups of three rings form inseparable triplets. Specifically, we focus on square lattices at which each node is embedded a loop which forms a Borromean link with pairs of its nearest neighbors. By mapping the Borromean link network onto a lattice representation, we investigate the fraction of occupied nodes required for a giant component, (the percolation threshold), the spectrum of topological links that would be released if the network were dissolved to varying degrees. We find that the percolation threshold of the Borromean square lattice occurs when approximately 60.75% of nodes are occupied, slightly higher than the 59.27% typical of a square lattice. Compared to the dissolution of Hopf-linked networks, a dissolved Borromean network will yield more isolated loops, and fewer isolated triplets per single loop. |
Wednesday, March 8, 2023 3:48PM - 4:00PM |
Q02.00003: No Free Lunch for Avoiding Clustering Vulnerabilities in Distributed Systems Pheerawich Chitnelawong, Andrei A Klishin, Greg Van Anders Emergent design failures are ubiquitous in multi-component, distributed systems. For example, leakage in microprocessors, congestion in airline networks, and outfit density in warships are design failures that emerge from the undesirable clustering of design elements. Here, we use methods of statistical physics to show that clustering vulnerabilities arise from generic interactions among competing degrees of freedom that describe the placement and connection of distributed system elements. We show that clustering vulnerabilities have multiple origins, including trade-offs between configurational and conformational entropy, and that avoiding clustering is only possible at the expense of introducing large variability in arrangement. Our results suggest that there is "no free lunch" for avoiding clustering vulnerabilities, but they can be quantified and managed via statistical physics approaches. |
Wednesday, March 8, 2023 4:00PM - 4:12PM |
Q02.00004: The journal Φ index and highly cited papers Manolis Antonoyiannakis The journal Φ index is a standardized scale-independent citation indicator. Since it is free from the scale dependence that plagues citation averages due to the skewness of citation distributions, it has been proposed [1] as a fairer indicator than the Journal Impact Factor for journal comparisons. But how well does the Φ index capture a journal's citation footprint with regard to highly cited papers? This question matters for research assessment, because of the increased difficulty for transformative, groundbreaking papers—many of which become highly cited eventually—to appear in high Impact Factor journals. To answer this question we analyze papers that were cited at the 99th percentile in their academic field per publication year, and were designated as "highly cited papers" by Clarivate Analytics' Essential Science Indicators. We compare journal rankings based on the number of highly cited papers, with Φ index and Impact Factor rankings. We find that Φ index rankings have a considerably higher similarity with rankings of highly cited papers. Thus, apart from removing the scale dependence of Impact Factor rankings, the Φ index is also a better descriptor of a journal's potential at the high end—in capturing highly cited papers. |
Wednesday, March 8, 2023 4:12PM - 4:24PM |
Q02.00005: Self-linking duplication-divergence model for protein interaction networks Leopold Bilder, Istvan A Kovacs Models of protein-protein interaction networks provide insights into the structure and evolutionary origin of these networks as well as how they behave. The standard duplication-divergence model simulates the degree distribution of real protein-protein interaction data accurately (1). However, this model does not include self-interactions, although the corresponding homodimers play a key role in biology. Here, we extend and analyze the duplication-divergence model with self-linking nodes. Copying a self-linking node leads to a self-linking child with probability σ that connects back to the parent node. This addition gives rise to higher average degree for self-linking nodes than non-self-linking nodes, as is seen in real protein-protein interaction data. Also, in real data the clustering coefficient generally increases as the fraction of self-linking neighbors on a node increases, another pattern captured by our model. Overall, due to the model’s sensitivity to initial conditions, self-links have a major impact on the network topology at all scales.
|
Wednesday, March 8, 2023 4:24PM - 4:36PM |
Q02.00006: A mutation-selection paradigm for the dynamics of biological networks Vito Dichio, Fabrizio De Vico Fallani
|
Wednesday, March 8, 2023 4:36PM - 4:48PM |
Q02.00007: Structure of brain connectome and contactome in fly, mouse, and human Anastasiya Salova, Istvan A Kovacs Analyzing networks of neurons is fundamental to understand the structure and function of the brain as a whole. However, it poses unique challenges, since even the simple questions of what the node positions are and what counts as an edge require careful consideration. Additionally, standard spatial network generative models are not applicable to neuronal networks since they can not take into account the interplay of the complex fractal neuron structure, spatial orientation, and biological wiring rules. |
Wednesday, March 8, 2023 4:48PM - 5:00PM |
Q02.00008: Complex quantum networks on a lattice of spins Ravi T Chepuri, Istvan A Kovacs Recent findings have suggested that a future quantum internet could benefit from network complexity, as characterized by a heavy-tailed degree distribution, small world property, etc. However, the impact of complexity on quantum communication networks is still not fully understood, and there are known challenges. Here, we present a model of complex quantum communication networks in the context of interacting quantum spin systems on a lattice. In particular, we consider the two-dimensional Random Transverse Ising Model (RTIM), which has a ground state that conveniently factorizes into independent GHZ spin clusters. We show that these GHZ clusters can act as quantum communication links between local regions of spins on the lattice. By appropriately choosing many local regions of spins from a broad size distribution, the GHZ clusters can connect the regions in a network which we show can have considerable network complexity. The resulting model of a complex quantum communication network is generalizable to other interacting quantum spin systems, and is amenable to further studies towards understanding the role of complexity in a future quantum internet. |
Wednesday, March 8, 2023 5:00PM - 5:12PM |
Q02.00009: How good is the overlap between two biological networks? Bingjie Hao, Istvan A Kovacs Unlike in physics where experimental results are compared to theoretical models, a biological signal is often detected by comparing to a negative benchmark. As an example, the signal content of biological networks (e.g. protein-protein interactions) is traditionally highlighted by showing significant overlap with a gold standard network. Significance is then assessed by comparison with a negative benchmark in which the gold standard network is randomized. Although a significant difference from the negative benchmark indicates the presence of signal, the low observed overlap between most biological networks questions the quality of these networks. Thus, a positive benchmark is needed to quantify how much a network agrees with a gold standard network. Here, we propose such a positive benchmark that captures a best-case scenario by creating an alternative network ensemble. The alternative networks contain only the links of the union of the two networks and preserve the original degree sequence, based on the maximum entropy framework. The negative and positive benchmarks together allow us to normalize the observed overlap of two networks. As an application, we compared the molecular and functional networks in yeast as well as in human, leading to an agreement network of networks for each organism. Our results show that most networks of a given organism agree well with each other despite the low observed overlap, providing insights for network data integration as well as assessing new network datasets. |
Wednesday, March 8, 2023 5:12PM - 5:24PM |
Q02.00010: Applying a network reconstruction method to reveal connectivity of in-vitro neuronal culture from measurements Emily S.C. Ching We discuss our application of a method for reconstructing directed networks from dynamics [Ching and Tam, Phys. Rev. E 95, 010301(R) (2017)] to reveal the directed links and synaptic weights of in-vitro cortical neuronal cultures from voltage measurements recorded by a multielectrode array. The voltage signal measured by each electrode after noise reduction is treated as the activity xi(t) of node i, i = 1, 2,..., N, of a neuronal network of N=4095 nodes. The reconstructed connectivity reproduces various reported features of cortical regions in rats and monkeys. The average synaptic strengths of excitatory incoming and outgoing links in the reconstruction are found to increase with the spiking activity recorded in the electrodes. Numerical simulations of a network of spiking neurons using the reconstructed effective connectivity can reproduce such dependence of the network features on firing rates and the long-tailed distribution of firing rate found in the multielectrode recordings. Our results thus support that the reconstructed connectivity can capture the general properties of synaptic connections and reveal relationships between network structure and dynamics of neuronal cultures. |
Wednesday, March 8, 2023 5:24PM - 5:36PM |
Q02.00011: Bifurcations in the Herd Immunity Threshold for Discrete-Time Models of Epidemic Spread Maximilian M Nguyen, Sinan A Ozbay, Bjarke F Nielsen, Simon A Levin We performed a complete sensitivity analysis of the herd immunity threshold for discrete-time SIR compartmental models with a static network structure. We find unexpectedly that these models violate classical intuition which holds that the herd immunity threshold should monotonically increase with the transmission parameter. We find the existence of bifurcations in the herd immunity threshold in the high transmission probability regime. The extent of these bifurcations is modulated by the graph heterogeneity, the recovery parameter, and the network size. We observe this behavior in both network- and differential equation-based models, suggesting this behavior is a universal feature of all discrete-time SIR models. This suggests careful attention is needed in selecting the assumptions on how to model time and heterogeneity in the standard epidemic models that are used. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700