Using AI to make sense of genetic data

Last year the Obama administration launched an ambitious ‘moonshot’ to make a serious dent in our understanding and treatment of cancer.

This has been followed up by the recent launch of the Deloitte X Prize, which aims to further support innovative solutions to the fight against cancer, with a $20 million prize fund to help things along.

A central challenge is that our ability to generate data has grown tremendously, but our ability to make sense of it has not managed to keep pace.  Genomic sequencing, for instance, has plummeted in price, whilst we can now generate our own data via wearable devices or mobile apps.

Sense making

Machine learning may be able to play a part, with a recent partnership announced between DNA sequencing company Illumina and IBM Watson a good example.

Executives at the company reveal that clients have struggled to know how to make best use of the data they increasingly have available to them.  They didn’t know which tests to run, or how to interpret the results.  That’s where Watson hopes to come in, as they are using their AI smarts to help interpret the data more effectively.

It’s being trained on raw genomic data at around 20 of the leading cancer institutions across the United States.  The training is designed to help Watson better identify genetic variants that are known to play a role in informing treatment options.

“We believe that, in the long term, genomics has the potential to impact health care on a scale that we have never seen before,” Illumina say. “But that only happens to the extent that we get genomics into the clinic.”

Related

Facebooktwitterredditpinterestlinkedinmail

Leave a Reply

Your email address will not be published. Required fields are marked *

Captcha loading...