The cost function is rapidly changing toward computation and storage; for instance, we have collected over 1PT of data from Dr. Michael Snyder over the years!
Storing 1PT of data on any major cloud provider (on a hot storage) would cost over $20K per month. Our team has been working on not only AI aspects of medical problems but also on delivering a new generation of computational and storage pipelines to significantly increase the scalability of precision medicine.
We recently published a paper, Hummingbird: Efficient Performance Prediction for Executing Genomic Applications in the Cloud, in collaboration with the VA’s Million Veterans Project on how to optimize computational pipelines. The result is an open-source tool that saves researchers up to 800% in cloud computing costs!
How are we doing this?
One of our unique capabilities has been recruiting computer scientists across the US and training them here at Stanford School of Medicine – last year we received over 900 applications.
