IBM Watson Health has expanded its partnership with the Broad Institute of MIT and Harvard to work on a project that will leverage genomics, clinical data and artificial intelligence (AI) to predict serious cardiovascular diseases.
The initiative is set to result in tools for doctors to gain better insights into the intrinsic possibility that an individual has a particular disease, thus allowing appropriate intervention to potentially reduce the risk.
During the three-year project, the team will use both population and hospital-based biobank data, genomic information and electronic health records to increase the predictive capability of polygenic scoring (genetic risk scoring).
The researchers intend to create algorithms that could identify and learn from trends in these datasets and point towards a potential predisposition to some health disorders.
The AI technology will be designed to generate models that will combine and analyse multiple genetic risk factors within a genome.
These models will also integrate existing health records and biomarkers, allowing more accurate prediction of the onset of heart conditions such as heart attacks, sudden cardiac death and atrial fibrillation.
IBM Watson Healthsenior vice-president John Kelly said: “We’ve built a deep expertise in applying AI to understand the complexities and meaning of immense amounts of data, such as genomics and health records.
“Our latest collaboration will combine these capabilities with clinical insights, and refine how technology can provide explainable and valuable insights to clinicians as they study and treat serious conditions such as cardiovascular disease.”
Insights and tools from the project will be made available to the research community. This will include methods of calculating a patient’s risk of developing common diseases by using genomic variants.
Broad Institute Cardiovascular Disease Initiative director Sekar Kathiresan said: “We’re excited to build upon the advances we’ve made in polygenic risk scoring utilising vast amount of genomic data
“By coupling clinical data with genomic data, there is an exceptional opportunity to make polygenic risk scoring more robust and powerful, and ultimately transformative for patient care. Such transformation could never happen without these kinds of partnerships.”
Previously, the partners initiated a five-year project in 2016 to enable the use of machine learning and genomics to better understand the reason why cancers become resistant to therapies.