This is how algorithms predict the spread of lung disease
OSNABRÜCK The Canadian disease forecasting service Bluedot has achieved great success: its algorithm warned of a virus outbreak in Wuhan, China, as early as December 31, 2019. Nine days earlier than the World Health Organization. “Digital technologies such as data analysis and artificial intelligence are already making an important contribution to the fight against the coronavirus epidemic worldwide,” emphasizes Bitkom Chief Executive Bernhard Rohleder in an interview with our editorial team. This includes, among other things, the analysis of patient data. They would also be used in some countries to match movement data to identify contact persons of infected people. For example, the trade magazine “Futurezone” reports on a project at the Technical University in Vienna, where research is currently being carried out on how the virus is now spreading further. One basic finding is that the number of personal contacts is the biggest factor in the spread. If everyone reduces the number of their own contacts by 25 percent, the peak of new cases would fall by 42 percent. If you even give up half of the contacts, it is even more than 70 percent.
Networking of clinics In Germany, digital processes are reaching their limits, says Rohleder. “On the one hand, because, for example, a lot of patient data cannot yet be collected and evaluated digitally. Many official processes are still in progress by fax and paper.” On the other hand, data protection restricts, for good reason, official access to personal data such as location history. Only recently, the Robert Koch Institute’s considerations for mobile phone tracking to identify contacts of those infected with the coronavirus have met with fierce criticism. A suggestion by the Scientific Director Paul Lukowicz of the German Research Center for Artificial Intelligence (DFKI) is a trend-setting one. In an interview, the expert said, “if you connect the admission reports of all the clinics in the country, Which drug could help against Covid-19? Means. if an algorithm would constantly check these reports for conspicuous accumulations, the epidemic of the new coronavirus would probably have been noticed long before it spread.” Artificial intelligence could also help with the knowledge of the possible spread of a disease: “AI, for example, can process the data of international flight movements. Scientists around the world are working hard to find suitable and thus contribute to the prediction of how quickly something can spread,” explains Lukowicz. The use and disclosure of passenger data is subject to strict rules in Europe.
This does not seem to be the case in Canada. Because the disease specialist Bluedot not only searches the Internet for corresponding the references in regional news in 65 languages, databases and reports on animal and plant diseases and official health warnings as well as in forums and blogs. The algorithm also incorporates airline ticket data. Bluedot also correctly predicted that the Wuhan virus would develop in the first few days of the towards Bangkok, Seoul, Taipei and Tokyo. Artificial intelligence (AI) can also be helpful in the search for drugs that prevent the airway infection Covid-19 caused by the Sars-CoV-2 coronavirus from becoming non-life-threatening and quickly subsides. Among other things, AI is used to scour the molecular variants of effective vaccines in the pharmaceutical databases and to check their effectiveness. This can significantly shorten the process. “Drugs against SarsCoV-2 are needed quickly,” emphasizes Rolf Hömke, spokesman for the Association of ResearchIng Pharmaceutical Companies (vfa). But a fundamental new development of drugs for the virus would take more than a decade. “That’s why the focus of research to combat the Corona epidemic is on re-functioning existing drugs.” Doctors and pharmaceutical companies are currently investigating whether drugs that are either already approved for the treatment of another disease (e.g. hepatitis C or influenza) or which have at least already been tested against another disease with humans, against the Virus help. The possible renaming of the drugs is also referred to as “repurposing”.
Summit and Sierra, the two fastest supercomputers in the world, are also currently anticipating the coronavirus: Summit has already tested 8,000 active ingredients. Of these, 77 had an effect by preventing the S protein of the coronavirus SARS-CoV-2 from docking host cells. The data is now evaluated. The most promising active ingredients are then retested in the laboratory. Scientists are now using the Sierra supercomputer to create the most accurate model of the coronavirus. This will later test antibodies that are intended to attack and neutralize the virus. Vaccine development Some drug candidates are already on the trail. Among them “Remdesivir”, which was originally developed as a drug against Ebola. First patients have already received it. And it will also be used in Germany as part of studies. In Göttingen, infection biologists from the German Primate Center found what they were looking for. They rely on a drug approved in Japan, which is administered in case of inflammation of the pancreas. The drug “Camstat Mesulate” blocks the enzyme TMPRSS2. This, in turn, is responsible for allowing the coronavirus to penetrate human lung cells. The drug is now to be examined in clinical trials for its effectiveness in the fight against Covid-19. Digital technologies are also making an unprecedented pace in the development of vaccines against the new coronavirus. But it also has limits: experts believe it is unlikely that vaccines will be available for mass vaccinations this year.