Abu Dhabi Biobank teams with AstraZeneca on AI disease prediction

Abu Dhabi Biobank has teamed up with AstraZeneca on AI disease prediction. The two sides announced the collaboration during the Abu Dhabi Future Health Summit 2026 on 22 October. Abu Dhabi Biobank is a joint initiative of the Department of Health – Abu Dhabi and M42. The partnership will combine biological samples, genomic data, clinical records and artificial intelligence to support earlier detection of disease and research into rare conditions. One part of the work is to test AstraZeneca’s MILTON tool on Type 2 diabetes and cardiovascular disease. MILTON uses routinely collected health data to look for patterns linked to future disease risk. AstraZeneca developed it on data from the UK Biobank. The Abu Dhabi project will check whether the same approach works for the local population. The company has not published a start date for the local testing or the number of records that will be analyzed first. Abu Dhabi Biobank holds more than five million biological samples. It also links those samples to genomic information from the Emirati Genome Program and to long-term clinical records. The partners say this combined set can help produce evidence that reflects the UAE population rather than data drawn mainly from other regions. No independent audit of the sample count is included in the announcement. Slavé Petrovski, vice president of the Centre of Genomic Research at AstraZeneca, said the collaboration will support the creation of UAE-specific evidence that can inform national health policy and clinical decisions. The partners have not listed the specific clinics or hospitals that will supply the first data sets, nor have they named a date by which the first results will be published. For a patient the practical change is still ahead. The announcement describes a research collaboration. It does not describe a tool that is already running in clinics or a change in how a doctor orders a test tomorrow. The value depends on whether the partners later show that the models identify risk earlier in the local population and that clinicians can act on the result. Other countries in the region have also built biobanks and genome programs. Saudi Arabia has a large national genome project. Egypt has expanded digital health records. Abu Dhabi Biobank’s partnership with AstraZeneca is one local step that uses an existing AI model developed elsewhere and tests it against regional samples and records. The comparison is between a tool already built on UK data and a new check of how well it performs on UAE data. The collaboration does not list any new funding amount or any change in the number of samples stored. It lists the data sources that will be combined and the two disease areas that will be examined first. A reader can see the intended scope. A reader cannot yet see the accuracy the model achieves on local records or the share of high-risk cases that clinicians later confirm. The partners have not said how the results will be shared with clinicians or whether any tool will be made available for routine use. The current step is to assess whether MILTON can identify patterns in the UAE data. That assessment is still ahead. AstraZeneca and Abu Dhabi Biobank have worked together before on precision medicine. The new announcement adds a specific test of an existing AI model on local data for two common conditions. It does not replace earlier work. It adds one measurable question: how well the model performs on UAE records. The announcement does not name the individual researchers who will run the analysis or the computing environment that will host the model. Those details would show how the work will be done. Without them the partnership remains a statement of intent. MILTON was first described in a 2024 paper that used UK Biobank data. The Abu Dhabi work is a separate check on a different population. Results from one biobank do not automatically transfer to another. The partners have not said whether they will publish the local results in a journal or only in an internal report. If the partners publish the performance of MILTON on the UAE data for diabetes and heart disease and the number of records analyzed, the partnership has a measurable follow-through. If the next update is only a further signing, the change for patients remains the same as the initial announcement. The figure missing is the local accuracy of the model.