Can artificial intelligence predict how sick you will get from COVID-19? Scientists think so | national

SAN DIEGO – A team of scientists in San Diego are harnessing artificial intelligence to understand why symptoms of COVID-19 can vary dramatically from person to person – information that could prove useful in the continued fight against coronavirus and future pandemics.

Researchers looked at publicly available data to see how other viruses alter genes that our cells turn on or off. Using this information, they found a set of genes turned on in a wide range of infections, including the novel coronavirus. These genes predicted whether someone would have a mild or severe case of COVID-19, and whether they were likely to have a long hospital stay.

A team led by the University of California at San Diego, joined by researchers from Scripps Research and the La Jolla Institute for Immunology, released the results on June 11. The study authors say their approach could help determine whether new treatments and vaccines are working.

“When the whole world was confronted with this pandemic, it took people several months to scramble” to understand the new virus, said Dr Pradipta Ghosh, cell biologist at UC San Diego and one of the authors. of the study. “I think we need more of this IT framework to guide us through panic states like this.”

The project began in March 2020, when Ghosh teamed up with UC San Diego computer scientist Debashis Sahoo to gain a better understanding of why the new coronavirus was causing few or no symptoms in some people while wreaking havoc on others. other.

There was only one problem: The new coronavirus was, well, new, which means there wasn’t a lot of data to be drawn.

So Sahoo and Ghosh took a different approach. They searched public databases and downloaded 45,000 samples from a wide range of viral infections, including Ebola, Zika, influenza, HIV, and hepatitis C virus, among others.

Their hope was to find a common pattern of response to these viruses, and that’s exactly what they saw: 166 genes that were constantly turned on during infection. Of this list, 20 genes typically separated patients with mild symptoms from those who became seriously ill.

The coronavirus was no exception. Sahoo and Ghosh say they identified this common viral response pattern long before testing it in samples from COVID-19 patients and infected cells, but the results have held up surprisingly well.

“It seemed to work in all of the datasets we were using,” Sahoo said. “It was hard to believe.”

They say their results show that respiratory failure in COVID-19 patients is the result of overwhelming inflammation that damages the airways and, over time, makes immune cells less efficient.

Stanford’s Purvesh Khatri isn’t surprised. His lab regularly uses computer algorithms and statistics to find patterns in large datasets of the immune response. In 2015, Khatri’s group discovered that respiratory viruses trigger a common response. And in April, they reported that this shared response also applied to a range of other viruses, including the novel coronavirus.

This makes sense, Khatri says, because researchers have long known that there are certain genes that the immune system turns on in response to virtually any viral infection.

“Overall, the idea is pretty solid,” Khatri said of the recent UC San Diego study. “Genes are all (the) usual suspects.”

Sahoo and Ghosh continue to test their findings in the new coronavirus data as it becomes available. They are particularly interested in long-haul COVID-19. Ghosh says they are already finding that people with prolonged symptoms of coronavirus have distinct patterns of genetic activation compared to those who have fully recovered. Think of it as a smoldering fire that won’t go out.

The researchers’ ultimate hope is not just to predict and understand serious illness, but to stop it. For example, they say, a doctor might give a patient different treatment if a blood sample suggests that he is likely to become more ill with his current treatment. Ghosh adds that the genetic model they see could help identify promising new treatments and vaccines against future pandemics based on which therapies prevent serious disease responses.

“In uncharted and uncharted territory, this provides us (…) with safeguards to start looking for, understanding (the virus), finding solutions, building better models and, finally, finding therapies.”