These researchers use artificial intelligence work better vaccine against influenza

The influenza virus is particularly nasty bug, thanks in no small part as it mutates rapidly. Each time the flu virus changes, it becomes harder for people to develop immunity. Our best external defense against influenza, the vaccine against influenza, based on the best expert assumptions on which flu strain or strains are most likely widely used during a particular flu season. Sometimes they are right – this season the vaccine has been shown to be relatively good match to the strains circulating at the time. But sometimes they are wrong, they can lead to severe flu season. And the influence can extract a huge toll, killed each year in the US alone, tens of thousands of people and causing billions of dollars in hospital costs, not to mention lost productivity. It is there to fight the flu a precise way? The Berg team, a pharmaceutical startup based in Boston, thinks so. In collaboration with the French pharmaceutical giant Sanofi flu, Mountain and hopefully, new ways will be better understood with artificial intelligence, and machine learning found to stop it. the Berg researchers exposed the data (such as changes in mRNA and protein and metabolite concentrations) of hundreds of patient samples for influenza vaccine, this information is carried out by a sophisticated algorithm of artificial intelligence. Here mountain team hopes to understand better meet the biological processes that elicit an effective immune response to the flu. Niven Narain, Mountain co-founder and CEO says, such an understanding could help create vaccines against more precise influence, and therefore more effective. It could also pave the way for personalized prevention of influenza, developed with people receive customized vaccines to induce a better response in their immune system. That would have chosen a radically different approach from that of the World Health Organization (WHO) annual task of collecting different flu samples from around the world, including in every shot tribes year. “The work is very important, because we, of course, more efficiency and protection against lack of influence,” says Narain. For him, that means to be “to understand the biology of the patient, who worked on the vaccine, so we that Biomarkers predict discover [and make] to make the process more efficient.”, Say other experts to influence specific vaccines next limit could be fighting the flu. “Although we have used for decades a vaccine, we really do not handle exactly what kind of response should a new vaccine to give us the best protection is to lead,” said Richard Webby, director of the WHO Collaborating Center for Health to influence the ecology studies in animals to the hospital looking for San Jewish children. He is in the elite group of health experts who advise WHO on which strains of flu vaccine each year should go. “Find all the efforts in the field are very important – I would say that is exactly what he does [Mountain],” says Webby, who is not connected to the mountain. Jacco Boon, an assistant professor of medicine specializing in virology influence at Washington University in St. Louis, says AI-like efforts like Berg’s work have already yielded new information about the human immune response to the flu. “What we can see, and in what quantity, in the course have become better all the time”, says Boon. Webby says that mountain is partnering with Sanofi reason for increased safety at the start. “The machine learning approach that allows them to process the amount of data and allows them to measure many components of the immune response,” he says. “And the collaboration with Sanofi potentially allow the number of people needed to get.” In better flu vaccines for patients lead to at least a few more years, Boon adds that, even if successful, Mountain work is unlikely. Other researchers are using AI try to get influenza from other angles. Roni Rosenfeld, a computer scientist at Carnegie Mellon University, is using machine learning, the influence prediction spread by an algorithm on data from doctors’ offices and hospitals, training and social media posts, Google search trends, and retail sales information, among other sources. Much of the challenge in working Rosenfeld is in understanding what seems an ongoing outbreak yet given the available data. “Unlike weather forecasts, which measures how time is currently not that difficult background influence forecast, you have the additional problem … is that the current state of the world to know is not so trivial,” he says, in other words , figure a substantial portion of the expected influence where the disease at a given time is available. “Much of what we do is the prognosis, aber, Jetzt casting ‘,” he says. The Centers for Disease Control and Prevention (CDC) has long been monitoring flu trends were leading, but in recent years has encouraged the groups to develop Rosenfeld forecasting techniques. And starting last year, the CDC began an aggregation of such laboratories published forecasts (including Rosenfeld) on-line, so that both doctors and citizens can see where the flu has spread and where it is likely to It will spread. Even with mountain work and similar efforts, we are still far from a universal flu vaccine, along the Holy Grail of influenza research. But that does not mean there is not room for many improvements on the way, experts say. “Who wants to go in this kitchen sink approach all studies in this field, tell us something that can be useful,” says Webby. “And ‘the transformation or not remains to be seen, but I think it’s absolutely worth it.”
Image copyright Jeff Green UIG / Getty Images