Context: The EPFL COVID-19 academic committee has given the green light to twenty research projects directly addressing the COVID-19 crisis. In their research, teams from all the Schools work on priority aspects of the pandemic: preventing or tracking contagion, cure and vaccines, diagnostics and testing, and policy and data. Several teams are working together with partner institutions in Switzerland.
Objective: to present a selection of 4-5 COVID-19 projects to the C4DT community and stir interest for collaboration
Format: 1.5h online workshop. 4-5 project presentations (10 min per project), followed by an open discussion
Target group: C4DT community
Date: Tuesday 16/06/2020 @ 15h-16h30
Selected projects (our suggestions based on relevant projects, which were approved by EPFL’s academic committee):
1. The COVID-19 pandemic: an economic impact. Presenter: Rafael Lalive.
The Coronavirus pandemic is an urgent and dire health challenge. At the heart of the difficult decisions taken by authorities in the fight against the disease is the tension between the need for radical and early preventive measures, on the one hand, and on the other the perceived necessity to minimize the economic and social costs of the disruptions imposed on society by the measures themselves. Teams of researchers affiliated with the E4S Center (EPFL-UNIL-IMD) explore several aspects of this tension between sanitary and economic concerns, particularly in view of being better prepared in the event of new pandemic outbreaks.
How countries reacted to the outbreak of the virus in their own populations differed substantially, giving rise to speculations about when is the best time to intervene. When choosing these measures, policy makers were aiming to improve Public Health while keeping the flow of goods and people in the country running. Countries like China, Singapore and Japan reacted fast and closed schools or locked people in their homes, while countries like Italy or Switzerland waited longer and limited movements of their population only once severe shortages in Public Health systems were foreseeable. Using data on measures taken by countries, and novel data on public opinion building on web-scraping techniques, we study the extent to which economic, political and sociodemographic aspects of countries predict the speed at which countries take measures, and how reacting fast vs slow affects health and economic outcomes. These analyses will help understand better the patterns of decisions and consequences in the world-wide Covid19 pandemic.
2. CoughVid: COVID-19 detection through cough recording app. Presenter: Tomas Teijeiro.
According to statistics published by the WHO, 67.7% of people who contract COVID-19 have a distinctive, dry cough. Researchers at EPFL’s Embedded Systems Laboratory quickly realised that artificial intelligence should be able to identify such a distinctive sound, and that this could provide an automated, centralised way to diagnose the disease on a portable platform. The CoughVid initiative aims to provide a free and simple diagnostic tool through a website, available here
3. Host genetics of life-threatening SARS-CoV-2 infection in previously healthy patients. Presenter: Jacques Fellay
The variability observed in the course of COVID-19 makes the existence of human genetic factors influencing the response to SARS-CoV-2 very likely. To search for genetic variants predisposing to the most severe clinical presentation of the disease, we sequence the genome and the blood transcriptome of selected COVID-19 patients younger than 50 without comorbidity, who require ventilation in an intensive care unit. The identification and characterization of genetic variants and expression profiles responsible for severe COVID-19 in otherwise healthy individuals will help uncover the genes and pathways that play a crucial role in viral pathogenesis and in anti-viral response, which might inform drug and vaccine development.
4. Privacy-preserving COVID-19 data sharing while respecting patient privacy. Presenter: Jean-Pierre Hubaux.
A proper understanding of COVID-19 requires accessing data that is collected around the planet. Global sharing of this data will face the well-known hurdle of data protection. In this project, we make use of MedCo, a system that enables computation on data without moving it or decrypting it. MedCo has already been installed in three Swiss University hospitals (read the news).
5. Forecasting Covid cases and deaths curves at the country/state level globally (replaces Understanding epidemic spreading) Presenter: Guillaume Obozinski
In collaboration with the Global Health Institute in Geneva, the Swiss Data Science Center put in place a dashboard providing analytics on Covid from more than 200 countries worldwide and for individual states in the US. The dashboard provides in particular forecasts on a horizon of a week for the evolution of the number of cases, the number of deaths and the evolution of the effective reproducing number. In this talk, I will present our modeling approach and the challenges associated with working with this kind of data.