Superagers and what they can tell us about dementia
What are the researchers trying to find out?
The research team based in Maastrict University, Kings College London and the University of Oxford want to discover the key factors that allow people to live well – without developing dementia – so that they can build a computer prediction model of successful cognitive ageing. Computer or 'in silico' models like this are important because they can be used in the testing of potential treatments early on the drug discovery process.
Creating a new resource
To be as useful as possible, these 'in silico' models need to be based on sufficiently large datasets – just one cohort's data would not be big enough for a reliable model. The first task for the researchers was to bring together all the data that has been collected on the longest-living people. They needed cohorts which have collected lots of different types of information on people who have been followed up into their 90s.
Using DPUK's cohort discovery tools – the cohort matrix and the cohort directory – the team identified eight DPUK cohorts that could be useful. They were approved to work with six of these and by combining them with the six EMIF cohorts, the team now has access to records of over 1500 'superagers' which they are processing in their secure area of the Data Portal's analysis environment.
The 'superagers dataset'
By bringing together the cognitive, lifestyle, genetic, biomedical, demographic and psychosocial records from over 1500 people aged 85 and over, the team have created a new database – the 'superagers' dataset – based on the data from the six DPUK cohorts. The initial summary descriptive data from these records was presented at the EMIF 2018 symposium. Following further data cleaning, this dataset will be made available to researchers worldwide to use in their analyses or 'in silico' experiments through the Data Portal.
New disease prediction models
When the team complete the processing of this complex new dataset, they will begin building multilevel statistical models to better understand brain health and disease in old age. A cohort profile paper is also planned for publication later this year.
The researchers in this team are based in three different organisations, located in the UK and the Netherlands: