VEO participants (and laboratories) were called into action on regional, national and international scales to work on COVID-19. In addition, some of the planned activities were re-focused in order to put scientific expertise within VEO in support of the pandemic response: VEO was involved in development of the first diagnostic PCR and serology assays for COVID-19 in the world; in seminal studies demonstrating the infection kinetics and pathogenesis of COVID-19 in comparison with MERS and SARS in animal models; in studies exploring potential animal reservoirs; in citizen science-based efforts to understand the impact of social media dynamics; in deployment of sewage testing to understand the population-level impact of the pandemic; and in modelling potential trajectories of the pandemic in different scenarios. Upon request by the European commission, VEO helped develop the COVID-19 Data Portal, and built bioinformatic workflows that allow processing of raw sequence data in a standardized manner, complementary to the genome sharing through GISAID. This has provided the largest ever publicly available pathogen genome and variant database open for re-use. The SARS-CoV-2 genomic data constitutes 25% of all globally available genomic data of pathogens, including influenza, polio, measles, foodborne bacteria, TB, and HIV.
Despite the delays due to the pandemic, the use-case scenario work packages progressed well, with some adjustments to the plans. The mosquito-borne work took the observations of new detections of West Nile Virus (WNV) in Northern Europe. In the recently completed reporting period, integrated analysis of data on land use, bird migration data, climate variables and field data showed that in fact there are distinct dispersal patterns of these viruses in Europe, driven by different drivers. Land use change was listed as one of the key drivers. As information on mosquito presence and abundance is a key component of models predicting dispersal, the Mosquito Alert app was used to collect such information through citizens. In line with this broader engagement, a data challenge was called, in which more than 1000 data scientists from 100 teams from all over the world worked on improving the current models for mosquito identification.
Building from the same approach, the global epizootic of highly pathogenic avian influenza viruses was tracked, reconstructing from genomic data coupled with information of (migratory) birds what the most likely route of dispersal had been. The ongoing analyses also show where hotspots are for further mixing of genes of this evolving pathogen.
As Greenland is important in the ecology of wild birds and is one of the most affected ecosystems due to climate change, we were interested to study its role in global dissemination of bird-borne and vector-borne pathogens. Two expeditions to Greenland were organized to look into exposures to avian influenza but also other pathogens in birds, mosquitos and mammals in Greenland. This confirmed evidence of exposure to H5 influenza, but also made clear how variable the climate is, with large differences in environmental suitability for bird breeding in the two years. This work also was used for a cross consortium initiative to address risks and benefits of biosurveillance studies.
This work has continued with the spill over of H5N1 avian influenza into livestock (cattle) in North America, which was unexpected and unprecedented and raised new questions about genomic markers predicting host range and mammalian adaptation. A new activity in VEO has been the start of exploratory studies that aim to predict phenotypes from pathogen genomes, building from historic datasets and using different machine-learning approaches.
Part of the silent epidemic scenario is the exploration of the use of wastewater surveillance as an early warning indicator. This approach moved into the spotlight during the pandemic, where wastewater surveillance was established across Europe, including through a Joint Research Centre (JRC) coordinated effort. VEO is exploring how this could be expanded to any pathogen by using metagenomic sequencing, which generates data of a size and complexity that currently is difficult to host in commonly used databases in public health. Following initial publications on antimicrobial resistance genes in untreated sewage from 101 countries, a deeper analysis focused on tools for mining the large fraction of any metagenomic dataset that is not a known pathogen or resistance gene. This dark matter mining requires highly advanced bioinformatic analyses, which has been piloted in sequential samples from four cities in Europe.
