Watson for Drug Discovery identifies proteins associated with cardiovascular disease

Capsule close-up.

Heart attack and stroke are 2 of the top 10 causes of death worldwide. Identifying the proteins involved in each of them is the beginning of a journey to understanding how these diseases develop. In a novel experiment presented at the American Heart Association Scientific Sessions on November 12, 2017, and published in a supplement of the journal Circulation, cardiology researchers tested Watson’s ability to rank proteins they knew were associated with cardiovascular disease. In the experiment, researchers gave Watson a list of 1,274 proteins associated with many diseases, some of which the researchers knew were associated with atherosclerosis and heart failure to see if Watson could identify them.

Watson used a training sample of proteins known to be associated with atherosclerosis and heart failure to rank the remaining proteins, by comparing literature fingerprints. Watson ranked the cardiovascular proteins significantly higher than non-cardiovascular proteins and its predictive ability improved with increased specificity of training on the cardiovascular disease phenotypes of coronary atherosclerosis and heart failure. These data demonstrate that Watson for Drug Discovery does have the ability to help researchers by classifying proteins associated with cardiovascular disease and that the predictive ability is honed with increasing specificity of disease phenotype.1

The next step will include analyzing blood samples taken from about 20,000 patients to evaluate whether Watson can successfully identify new proteins for researchers that have not previously been associated with heart failure and atherosclerosis. Congratulations to Scott Spangler and Alix Lacoste for their work in demonstrating Watson’s capabilities in cardiovascular disease discovery.

1. Christian T Ruff, Alix Lacoste, Francesco Nordio, Christina L Fanola, Michael G Silverman, Elenee Argentinis, Scott Spangler, Marc S Sabatine. Classification of Cardiovascular Proteins Involved in Coronary Atherosclerosis and Heart Failure Using Watson’s Cognitive Computing Technology. Circulation. 2017;136:A16678