When we move we carry many things with us: clothes, mobile phones, ideas, money… and also pathogens! In this work we construct the largest network of transfers of patients among US hospitals. We show that this transfer network actually correlates with cases of C.diff. infections.
J Fernández-Gracia, JP Onnela, ML Barnett, VM Eguíluz, NA Christakis
Scientific reports 7 (1), 1-9
One of the things that the pandemic has done to our world is that the scientific community has turned many of their efforts to understand and try to control it from several disciplines. This extraordinary worldwide collaboration, together with the modern information and communication technologies, efforts to share the data and modern machine learning and bioinformatics techniques have given us the ability to observe the evolution of the different strains of the virus in almost real time. Here we use genetic data on SARS-COV2 isolates to test results of evolutionary theories in real time. In particular we show the approximate adherence to Yule’s model of the evolution of the different strains and also show that the architecture of the phylogenetic trees representing the families of strains is neither completely balanced, nor completely imbalanced, as seen in earlier studies. We also find that the variants of concern have shown a higher diversification in the RBD domain, which is responsible for binding with human cells, and thus a determinant for the infectiousness of the strain.
CM Duarte, DI Ketcheson, VM Eguíluz, S Agustí, J Fernández-Gracia, ...
Scientific reports 12 (1), 1-8
There are many ecosystems in which mutualistic relations, the ones that are beneficial for both species in interaction, are crucial for their survival. Piotr Kropotkin noticed very early that mutual aid had to be an important factor in evolution. Examples of this kind of ecosystem are the ones in which we have insects pollinating plants. The structure of these ecosystems is captured by bipartite networks and research has already shown that there are several structural characteristics that seem to pervade in them and are crucial for their stability and robustness. Here we use data on interactions between plants and pollinators in Mallorca and show how the progressive extinction of different species under different assumptions changes the characteristics of the network.
S Sheykhali, J Fernández-Gracia, A Traveset, M Ziegler, CR Voolstra, ...
Scientific reports 10 (1), 1-12
In our life we are embedded in different networks - our friendships, our relatives, the hierarchies at the workplace or the transport networks we use in space. This networks are intertwined and their characteristics affect each other in different feedback loops. The study of them in isolation is useful, but in order to advance we need higher order structures. Take for example your network of acquaintances. They are probably spread in space, but , in order to form, you have mostly met in the same place with most of your friends. Here we propose a model for the embedding of one network into another, where the largest network is thought as a social network and the other is a spatial network. Using random walks with only one parameter - the stopping probability - we are able to create embeddings of the social network in space that range from very local (all your friends are near) to very delocalized (where your friends could be very far away).
J Fernández-Gracia, JP OnnelaScientific reports9 (1), 1-7
The COVID-19 pandemic hit us hard all over the world. Islands are a suitable model for the study of their dynamics due to their quasi-isolation (even more during lockdown). During the first wave we studied the dynamics of the epidemics in the balearic islands and explored different scenarios for the risk of a second wave. The model we used is a typical SEIR model in a metapopulation context with recurrent mobility of agents. We were able to assess not only the risk of a secondary wave, but also the parameters of the model and the fraction of infected individuals that had gone unseen by the health monitoring system.
VM Eguíluz, J Fernández-Gracia, JP Rodríguez, JM Pericàs, C Melián
Frontiers in medicine 7, 905
Elections are crucial in democracy and their results form from different mechanisms from the individual to the collective. In this paper we show how a simple model of opinion competition, the Voter Model, with a realistic social context for the agents, based on the spatial population distribution and the commuting patterns of the individuals, is capable of reproducing statistical features of election results. In particular the model is able to capture the stationarity of the dispersion of vote-share distributions and the long range logarithmic decay in correlations of vote-shares.
J Fernández-Gracia, K Suchecki, JJ Ramasco, M San Miguel, VM Eguíluz
Physical review letters 112 (15), 158701