Edinburgh researchers have made a ‘significant breakthrough’ in predicting patients at risk of disease, including Alzheimer’s dementia, type 2 diabetes and heart disease.
University of Edinburgh experts teamed up with a data consultancy and a US biotech firm to determine patients’ risk factors from their genetic data.
They used cutting-edge machine learning to analyse vast amounts of medical data from the UK Biobank – a national repository for genetic data based in Stockport.
In a first of its kind study, published in Nature Aging, the researchers were able to identify protein patterns, also known as protein signatures, that are linked to the risk of diseases.
This allowed the researchers to accurately predict a person’s risk of disease up to 10 years before diagnosis. The research also showed how results from currently used patient blood tests could be compared with new protein patterns discovered in Biobank data.
In theory this would allow clinicians to detect the possibility of a particular disease developing later in life. If tests show a patient is at higher risk, there will be more time to proactively plan and take preventative measures to improve the eventual patient outcome.
Analysing data from a randomised set of almost 50,000 individuals who had a blood sample taken between 2006 and 2010, the study was able to improve the prediction for disease outcomes diagnosed up to 15 years after the initial blood sample was taken.
GDF15, a marker of inflammation, was among some of the proteins under investigation and was found to be linked with almost half (11 out of 23) of the diseases being studied, including both Alzheimer’s and vascular dementia, heart disease, liver disease, type 2 diabetes and all-cause mortality.
Dr Danni Gadd, the first author of the study, said: “Our research represents a promising step forward in risk prediction. It’s encouraging to see how much potential there is from a single blood sample that allow us to predict a range of disease outcomes. Being able to detect early warning signs for a broad set of conditions may lead to opportunities for early intervention and prevention, marking a significant moment for the healthcare industry.”
One of the study’s principal investigators, Dr Chris Foley, managing director and chief scientist of Optima Partners, said: “More work is still needed to convert these findings for practical use in clinical settings. However, our discoveries set strong foundations for the inclusion of new risk prediction signatures to shed a light on possible pathways and mechanisms that underlie diseases. Pattern recognition like this would not be possible without modern machine learning technology and its capacity to analyse data at this scale and will in turn allow us to address some of the most pressing healthcare challenges of our time.”
The study is published in the journal Nature Aging and was conducted by researchers from the Optima Partners, Biogen and the University of Edinburgh.
