Abstracts
Suppression of COVID-19 outbreak in the municipality of Vo’, Italy
Andrea Crisanti
University of Padova, Italy
ABSTRACT: On the 21st of February 2020 a resident of the municipality of Vo’, a small town near Padua, died of pneumonia due to SARS-CoV-2 infection1. This was the first COVID-19 death detected in Italy since the emergence of SARS-CoV-2 in the Chinese city of Wuhan, Hubei province2. In response, the regional authorities imposed the lockdown of the whole municipality for 14 days3. We collected information on the demography, clinical presentation, hospitalization, contact network and presence of SARS-CoV-2 infection in nasopharyngeal swabs for 85.9% and 71.5% of the population of Vo’ at two consecutive time points. On the first survey, which was conducted around the time the town lockdown started, we found a prevalence of infection of 2.6% (95% confidence interval (CI) 2.1-3.3%). On the second survey, which was conducted at the end of the lockdown, we found a prevalence of 1.2% (95% CI 0.8-1.8%). Notably, 43.2% (95% CI 32.2-54.7%) of the confirmed SARSCoV-2 infections detected across the two surveys were asymptomatic. The mean serial interval was 6.9 days (95% CI 2.6-13.4). We found no statistically significant difference in the viral load (as measured by genome equivalents inferred from cycle threshold data) of symptomatic versus asymptomatic infections (p-values 0.6 and 0.2 for E and RdRp genes, respectively, Exact Wilcoxon-Mann-Whitney test). Contact tracing of the newly infected cases and transmission chain reconstruction revealed that most new infections in the second survey were infected in the community before the lockdown or from asymptomatic infections living in the same household. This study sheds new light on the frequency of asymptomatic SARS-CoV-2 infection and their infectivity (as measured by the viral load) and provides new insights into its transmission dynamics, the duration of viral load detectability and the efficacy of the implemented control measures.
Data-driven modeling of COVID-19 pandemic
Yamir Moreno
University of Zaragoza, Spain
ABSTRACT: The new Coronavirus disease 2019 (COVID-19) has forced an unprecedented response from the authorities, first from the government of China and the World Health Organization, and later from many more countries as the disease spread worldwide. Despite the adoption of drastic measures, the pandemic is still ongoing worldwide, and surges of infections are being observed in more than 188 countries. Due to the lack of new specific pharmaceutical interventions or vaccines, the extent to which the adopted non-pharmaceutical interventions would be effective in the long term remains open. Here we present results from simulations using data-driven models tailored to mobility data from China, Spain, and the U.S. The models are used to estimate the effectiveness of customary public interventions on the spread of COVID-19 in these locations. Our main findings support incentivizing the adoption of actions that reduce the transmissibility of the disease as well as those aimed at improving the efficacy of early detection and isolation of newly symptomatic individuals. This highlights that having a coordinated response system could be key for the containment of the spread of COVID19 and its possible eradication at the lowest possible cost.
Predicting epidemic spreading and synchronization in complex networks
Francisco A. Rodrigues
University of Sao Paulo, Brazil
ABSTRACT: One of the most fundamental problems in Network Science is to understand how dynamical processes are influenced by the network organization. For instance, if we can understand how patterns of connections between coupled oscillators influence the evolution of the synchronous state, then we can change the network topology to control the level of synchronization of power grids and electronic circuits. Despite this fundamental importance, predicting the dynamical variable associated to a dynamical process, as the state of an oscillator or the outbreak size, from the network structure is a very complicated task in structured networks due to the presence of non-trivial patterns of connections, nonlinear effects and correlations between variables. However, we can gain some insights about the influence of network structure on dynamical processes by using mean-field approximation and machine learning algorithm. In this talk, we will show how the critical threshold for the emergence of synchronization and the critical probability for the occurrence of the endemic state depend on network properties. Moreover, we will discuss machine learning methods can be used to predict the outbreak size starting from a single node and the state of Kuramoto oscillators. The current challenges in Network Science and some possible ideas for future research will also be discussed in our talk.
Interpretable deep learning of label-free live cell images uncovers functional hallmarks of highly-metastatic melanoma
Assaf Zaritsky
Ben-Gurion University of the Negev, Israel
ABSTRACT: Deep learning has emerged as a powerful technique to identify hidden patterns in complex cell imaging data, but is criticized for the lack of insights it provides on the machine’s prediction. Here, we demonstrate that a generative adversarial neural network captures subtle details of cell appearance that allow classification of melanoma metastatic efficiency of patient-derived xenograft models that reflect clinical outcome. We used the network to generate “in-silico” cell images that amplified the cellular features critical for the classification. These images unveiled pseudopodial extensions and increased light scattering as functional hallmarks of metastatic cells. We validated this interpretation using live cells spontaneously transitioning between states indicative of low and high metastatic efficiency. Together, this data demonstrates how the application of Artificial Intelligence can support the identification of processes that are essential for the execution of complex integrated cell functions but are too subtle to be identified by a human expert.
How statistics of complex systems appears in non-equilibrium systems
Stefan Thurner
Medical University Vienna, Austria
ABSTRACT: Driven systems abound in the world of complex systems. We show that the underlying statistics of these systems can be undserstood by "sample space reducing" (SSR) processes that offer an intuitive understanding of the origin and ubiquity of power-laws in countless complex systems. We show that SSR processes exhibit a wide range of statistical diversity. We offer a way to implement the details of a driven system to arrive at practically any distribution function. Depending of the details of the driving process, we can be naturally derive Zipf’s law, exact power-laws, exponential, Gamma, normal, Weibull, Gompertz and Tsallis-Pareto distributions. We discuss the areas of applications of SRR processes that range from fragmentation processes, language formation, cascading processes and search processes in general.
Hypergraphs for analysing protein interaction networks
Florian Klimm
University of Cambridge and Imperial College London, UK
ABSTRACT: Protein-protein interactions are crucial in many biological pathways and facilitate cellular function. Investigating these interactions as a graph of pairwise interactions can help to gain a systemic understanding of cellular processes. It is known, however, that proteins interact with each other not exclusively in pairs but also in polyadic interactions and they can form multiprotein complexes, which are stable interactions between multiple proteins. In this manuscript, we use hypergraphs to investigate multiprotein complex data. We investigate two random null models to test which hypergraph properties occur as a consequence of constraints, such as the size and the number of multiprotein complexes. We find that assortativity, the number of connected components, and clustering differ from the data to these null models. Our main finding is that projecting a hypergraph of polyadic interactions onto a graph of pairwise interactions leads to the identification of different proteins as hubs than the hyper-graph. We find in our data set that the hypergraph degree is a more accurate predictor for gene-essentiality than the degree in the pairwise graph. We find that analysing a hypergraph as a pairwise graph drastically changes the distribution of the local clustering coefficient. Furthermore, using a pairwise interaction representing multiprotein complex data may lead to a spurious hierarchical structure, which is not observed in the hypergraph. Hence, we illustrate that hypergraphs can be more suitable than pairwise graphs for the analysis of multiprotein complex data.
Network methods for nonlinear time series analysis
Michael Small
University of Western Australia, Australia
ABSTRACT: Various network based methods exist for representing the structure and dynamical behaviour of dynamical systems when observed through a single scalar time series. These methods essentially provide an alternative to the usual delay-embedding approach to reconstruct dynamics from time series. However, whereas delay reconstructions are guaranteed to yield embeddings that generically preserve properties of the underlying dynamics, results for network methods are weaker. Nonetheless, we will discuss various alternative network construction techniques and how they can be employed to both reconstruct and then dynamics from time series.
From noisy point clouds to complete ear models: an unsupervised pipeline for application in the prosthetic industry
Filipa Valdeira
University of Milan
ABSTRACT: Ears are a particularly difficult region of the human face to model, not only due to the non-rigid deformations existing between shapes but also to the challenges in processing the retrieved data. The first step towards obtaining a good model is to have complete scans in correspondence, but these usually present a higher amount of occlusions, noise and outliers when compared to most face regions, thus requiring a specific procedure. Therefore, we present a complete pipeline taking as input unordered 3D point clouds with the aforementioned problems, and producing as output a dataset in correspondence, with completion of the missing data. We go over different approaches for registration and different paths for shape completion, comparing their performance on our data.
Biological Networks Regulating Cell Fate Choice are Minimally Frustated
David A. Kessler
Bar Ilan University, Israel
ABSTRACT: Characterization of the differences between biological and random networks can reveal the design principles that enable the robust realization of crucial biological functions including the establishment of dierent cell types. Previous studies, focusing on identifying topological features that are present in biological networks but not in random networks, have, however, provided few functional insights. We use a Boolean modeling framework and ideas from spin glass literature to identify functional dierences between five real biological networks and random networks with similar topological features. We show that minimal frustration is a fundamental property that allows biological networks to robustly establish cell types and regulate cell fate choice, and that this property can emerge in complex networks via Darwinian evolution. The study also provides clues regarding how the regulation of cell fate choice can go awry in a disease like cancer and lead to the emergence of aberrant cell types.
How do cancer cells make decisions during metastasis? (Causes & consequences of multistability in cancer metastasis )
Mohit K. Jolly
Centre for BioSystems Science and Engineering, Indian Institute of Science, Bangalore, India
ABSTRACT: Metastasis – the spread of cancer cells from one organ to another – causes above 90% of all cancer-related deaths. Despite extensive ongoing efforts in cancer genomics, no unique genetic or mutational signature has emerged for metastasis. However, a hallmark that has been observed in metastasis is adaptability or phenotypic plasticity – the ability of a cell to reversibly switch among different phenotypes in response to various internal or external stimuli. This talk will describe how mechanism-based mathematical models can help (a) identify how cancer cells can leverage this plasticity to drive cancer metastasis, (b) interpret confounding experimental and clinical data, and (c) guide the next set of crucial in vitro and in vivo experiments. Collectively, our work highlights how an iterative crosstalk between mathematical modeling and experiments can both generate novel insights into the emergent nonlinear dynamics of cellular plasticity and uncover previously unknown accelerators of metastasis.
African American Urban Health as a Complex Adaptive System- a call for Urban Bioethics and Complexity Ethics
Michele Battle-Fisher
Equitas Health Institute, Columbus, OH
ABSTRACT: The United States has gross health disparities that further marginalized populations already under immense social, financial, physical and emotional stressors. The percolating attention to health disparities brought about by COVID-19 and Black Lives Matter bears the question whether the current “system” produces ethical, equitable health for the urban African American community in the U.S. I argue that a paradigm shift that connects complexity thinking when used as scientific principle and as applied practice supersedes the gold standard of causal inferences when capturing social complexity. The Urban Bioethics framework accounts for the specific needs of the city’s inhabitants and advantages resources through full engagement of members of that community. Urban Bioethics as Complex Adaptive Systems and Complexity Ethics are material to unravel systemic, urban health inequities.
Mathematical morphology for the detection of face emotions based on facial expressions evolution from videos
Rongjiao Ji
University of Milan, Italy
ABSTRACT: Human facial expressions, which correspond to the simultaneous contraction or release of a set of facial muscles, convey informative but subtle messages, while human emotions are related and recognizable by specific facial expressions. From the perspectives of mathematics and computer graphics, we intend to give a realistic description of facial expressions under specific emotions, by considering facial action units data (the level of the engagement degree for a specific set of facial muscles) extracted from videos where actors or volunteers have been asked to represent a specific emotion. The goals of this work are to identify the main expressed emotion or quantify the mixture of emotions from action unit curves, and to describe mathematically latent patterns of face expression evolution (i.e. facial action units) common to specific human emotions. These goals imply general mathematical research focused on the development of a new model extraction algorithm by exploiting functional statistics and topological data analysis for multivariate cases.
Darkness Visible: Complexity and AI in Cosmological Experiments
Ofer Lahav
University College London, UK
ABSTRACT: The talk will start by summarising the status of large galaxy surveys and “tensions" in the standard cosmological Lambda-Cold-Dark-Matter model. We will then describe the important role of Artificial Intelligence and Machine Learning in analysing the next generation of surveys of billions of objects, and the training of the next generation of physicists for the big data challenges.