COVID-19’s effects answered by AI

Artificial intelligence has been utilised to recognise metabolic markers of COVID-19, a discovery potentially leading to better understanding and treatments for those who suffer from symptoms months after diagnosis.
Scientists from Surrey University collected blood samples of hospital patients, finding that COVID-19 changes people’s metabolism. The researchers observed that the second wave disrupted metabolites differently from the first, showing that the effects of COVID-19 changed over time.
Most patients’ metabolites returned to normal levels once they recovered from the virus, but a small number continued to be disrupted after infection for several months.
The author of the study from Surrey University, Dr Holly-May Lewis, stated: “It is thought that around two million people suffer symptoms of COVID-19 a month after infection, and eight hundred thousand people still experience symptoms a year later. So, it’s clear that this virus will be with us for some time and, therefore, the scientific community is duty-bound to better understand COVID-19 and why, for some, symptoms seem to linger longer than average.”
The study examined the blood samples of one hundred and sixty-four hospital patients – one hundred and twenty-three with COVID-19 and the rest without – across the first two waves of infection. Using an AI model with ninety-one per cent accuracy, the team identified six metabolites that can be used to recognise COVID-19.
The corresponding author of the study from Surrey University, Melanie Bailey, stated: “To our knowledge, this is the first time it has been demonstrated that COVID-19 is now affecting patients’ metabolism differently than it did in the initial wave – which we think is due to emerging variants. It is known that different COVID-19 variants have various associated symptoms, so it makes sense that this would relate to changes in blood chemistry. With this in mind, we then deployed artificial intelligence to identify biomarkers characteristic of COVID-19, regardless of the COVID-19 wave. This can help us better understand the limitations of the treatments being given and help to make better ones.”