Comparing transmission reconstruction models with Mycobacterium tuberculosis whole genome sequence data

Pathogen genomic epidemiology is now routinely used worldwide to interrogate infectious disease dynamics. Multiple computational tools that reconstruct transmission networks by coupling genomic data with epidemiological modelling have been developed. The resulting inferences are often used to inform outbreak investigations, yet to date, the performance of these transmission reconstruction tools has not been compared specifically for tuberculosis, a disease process with complex epidemiology that includes variable latency periods and within-host heterogeneity. Here, we carried out a systematic comparison of seven publicly available transmission reconstruction tools, evaluating their accuracy in predicting transmission events in both simulated and real-world Mycobacterium tuberculosis outbreaks. No tool was able to fully resolve transmission networks, though both the single-tree and multi-tree input implementations of TransPhylo identified the most epidemiologically supported transmission events and the fewest false positive links. We observed a high degree of variability in the transmission networks inferred by each approach. Our findings may inform the choice of tools in future tuberculosis transmission analyses and underscore the need for caution when interpreting transmission networks produced using probabilistic approaches.

Comparing human and model-based forecasts of COVID-19 in Germany and Poland

1AbstractForecasts based on epidemiological modelling have played an important role in shaping public policy throughout the COVID-19 pandemic. This modelling combines knowledge about infectious disease dynamics with the subjective opinion of the researcher who develops and refines the model and often also adjusts model outputs. Developing a forecast model is difficult, resource- and time-consuming. It is therefore worth asking what modelling is able to add beyond the subjective opinion of the researcher alone. To investigate this, we analysed different real-time forecasts of cases of and deaths from COVID-19 in Germany and Poland over a 1-4 week horizon submitted to the German and Polish Forecast Hub. We compared crowd forecasts elicited from researchers and volunteers, against a) forecasts from two semi-mechanistic models based on common epidemiological assumptions and b) the ensemble of all other models submitted to the Forecast Hub. We found crowd forecasts, despite being overconfident, to outperform all other methods across all forecast horizons when forecasting cases (weighted interval score relative to the Hub ensemble 2 weeks ahead: 0.89). Forecasts based on computational models performed comparably better when predicting deaths (rel. WIS 1.26), suggesting that epidemiological modelling and human judgement can complement each other in important ways.

The seasonality of three childhood infections in a pre-industrial society without schools

Background: The burden of many infectious diseases varies seasonally and a better understanding of the drivers of infectious disease seasonality would help to improve public health interventions. For directly transmitted highly-immunizing childhood infections, the leading hypothesis is that seasonality is strongly driven by social gatherings imposed by schools, with maxima and minima during school terms and holidays respectively. However, we currently have a poor understanding of the seasonality of childhood infections in societies without schools and whether these are driven by human social gatherings. Here, we used unique nationwide data consisting of >40 epidemics over 100 years in 18th and 19th century Finland, an agricultural pre-health care society without schools, to (i) quantify the seasonality of three easily identifiable childhood infections, smallpox, pertussis and measles and (ii) test the extent to which seasonality of these diseases is driven by seasonal social gatherings. Methods: We quantified the seasonality of transmission using time series Suscpetibel-Infected-Recovery models, wavelet analyses and general additive mixed models.Results: We found that all three infections were seasonal and the seasonality patterns differed from those in industrialized societies with schools. Smallpox and measles showed high transmission in the first half of the year, but we could not associate this with seasonal human gatherings events. For pertussis, however, transmission was higher during social gathering events such as New Year and Easter.Conclusions: Our results show that the seasonality of childhood infections is more variable than previously described in other populations and indicate a pathogen-specific role of human social aggregation in driving the infectious disease dynamics.Funding: Academy of Finland (278751, 292368), Nordforsk (104910), the Ehrnrooth Foundation, the Finnish Cultural Foundation, the University of Turku Foundation and the Doctoral Programme in Biology, Geography and Geology, University of Turku.

A model for the co-evolution of dynamic social networks and infectious disease dynamics

Recent research shows an increasing interest in the interplay of social networks and infectious diseases. Many studies either neglect explicit changes in health behavior or consider networks to be static, despite empirical evidence that people seek to distance themselves from diseases in social networks. We propose an adaptable steppingstone model that integrates theories of social network formation from sociology, risk perception from health psychology, and infectious diseases from epidemiology. We argue that networking behavior in the context of infectious diseases can be described as a trade-off between the benefits, efforts, and potential harm a connection creates. Agent-based simulations of a specific model case show that: (i) high (perceived) health risks create strong social distancing, thus resulting in low epidemic sizes; (ii) small changes in health behavior can be decisive for whether the outbreak of a disease turns into an epidemic or not; (iii) high benefits for social connections create more ties per agent, providing large numbers of potential transmission routes and opportunities for the disease to travel faster, and (iv) higher costs of maintaining ties with infected others reduce final size of epidemics only when benefits of indirect ties are relatively low. These findings suggest a complex interplay between social network, health behavior, and infectious disease dynamics. Furthermore, they contribute to solving the issue that neglect of explicit health behavior in models of disease spread may create mismatches between observed transmissibility and epidemic sizes of model predictions.

Predicting partially observed processes on temporal networks by Dynamics-Aware Node Embeddings (DyANE)

Low-dimensional vector representations of network nodes have proven successful to feed graph data to machine learning algorithms and to improve performance across diverse tasks. Most of the embedding techniques, however, have been developed with the goal of achieving dense, low-dimensional encoding of network structure and patterns. Here, we present a node embedding technique aimed at providing low-dimensional feature vectors that are informative of dynamical processes occurring over temporal networks – rather than of the network structure itself – with the goal of enabling prediction tasks related to the evolution and outcome of these processes. We achieve this by using a lossless modified supra-adjacency representation of temporal networks and building on standard embedding techniques for static graphs based on random walks. We show that the resulting embedding vectors are useful for prediction tasks related to paradigmatic dynamical processes, namely epidemic spreading over empirical temporal networks. In particular, we illustrate the performance of our approach for the prediction of nodes’ epidemic states in single instances of a spreading process. We show how framing this task as a supervised multi-label classification task on the embedding vectors allows us to estimate the temporal evolution of the entire system from a partial sampling of nodes at random times, with potential impact for nowcasting infectious disease dynamics.

Comparing metapopulation dynamics of infectious diseases under different models of human movement

Newly available datasets present exciting opportunities to investigate how human population movement contributes to the spread of infectious diseases across large geographical distances. It is now possible to construct realistic models of infectious disease dynamics for the purposes of understanding global-scale epidemics. Nevertheless, a remaining unanswered question is how best to leverage the new data to parameterize models of movement, and whether one’s choice of movement model impacts modeled disease outcomes. We adapt three well-studied models of infectious disease dynamics, the susceptible–infected–recovered model, the susceptible–infected–susceptible model, and the Ross–Macdonald model, to incorporate either of two candidate movement models. We describe the effect that the choice of movement model has on each disease model’s results, finding that in all cases, there are parameter regimes where choosing one movement model instead of another has a profound impact on epidemiological outcomes. We further demonstrate the importance of choosing an appropriate movement model using the applied case of malaria transmission and importation on Bioko Island, Equatorial Guinea, finding that one model produces intelligible predictions of R0, whereas the other produces nonsensical results.

Haplotype Explorer: an infection cluster visualization tool for spatiotemporal dissection of the COVID-19 pandemic

Summary Many of software for network visualization are available, but existing software have not been optimized to infection cluster visualization, especially the current worldwide invasion of COVID-19 since 2019. To reach the spatiotemporal understanding of epidemics, we have developed Haplotype Explorer. In Haplotype Explorer, users can explore the network interactively with metadata like accession number, locations, and collection dates. Time dependent transition of the network can be exported as continuous sections for making a movie. Here, we introduce features and products of Haplotype Explorer, demonstrating time-dependent snapshots and a movie of haplotype networks inferred from total of 4,282 SARS-CoV-2 genomes. Abstract The worldwide eruption of COVID-19 that began in Wuhan, China in late 2019 reached 10 million cases by late June 2020. In order to understand the epidemiological landscape of the COVID-19 pandemic, many studies have attempted to elucidate phylogenetic relationships between collected viral genome sequences using haplotype networks. However, currently available applications for network visualization are not suited to understand the COVID-19 epidemic spatiotemporally due to functional limitations, that motivated us to develop Haplotype Explorer, an intuitive tool for visualizing and exploring haplotype networks. Haplotype Explorer enables to dissect epidemiological consequences via interactive node filters and provides the perspective on infectious disease dynamics depend on regions and time, such as introduction, outbreak, expansion, and containment. Here, we demonstrate the effectiveness of Haplotype Explorer by showing features and an example of visualization. The demo using SARS-CoV-2 genomes are available at https://github.com/TKSjp/HaplotypeExplorer/blob/master/Example/. There are several examples using SARS-CoV-2 genomes and Dengue virus serotype 1 E-genes sequence.