Microsoft Genomics service launched to accelerate healthcare innovation

By Elizabeth Robinson on 02 March 2018

Microsoft has launched a new cloud-based Genomics service. Leveraging Microsoft Azure, the service will enable researchers from around the world to collaborate in the search for cures to paediatric cancers.

In a recent blog post, John Roach, technical writer and editor at Microsoft, discusses the Microsoft Genomics service and how artificial intelligence (AI) and cloud computing can be used to help store and share data more efficiently.

“Variants are what make individuals unique,” said Roach. “They are markers for traits ranging from physical attributes to disease susceptibility. Variants are also the fuel for so-called genome-wide association studies that allow researchers to home in on what the variants mean. The more genomic data researchers can access and analyse, the more precisely they can tease apart the complexity of biology and make progress toward curing diseases such as cancer.”

Jinghui Zhang, computational biologist at St Jude Children’s Research Hospital, and her team worked with the Microsoft researchers on the development of the genome alignment and variant calling pipeline in partnership with DNAnexus, a secure, Azure-based platform for managing genomic data.

The genomic data analysed and stored in the cloud is the foundation for a data-sharing platform that the research hospital is building with DNAnexus and Microsoft.

“For us to be able to test on real-world data like that and be able to work side-by-side with these teams was an amazing opportunity,” said Geralyn Miller, director at Microsoft Genomics AI and Research.

A single human genome takes up 100 gigabytes of storage space. As more and more genomes are sequenced, storage needs will grow from gigabytes to petabytes to exabytes. By 2025, an estimated 40 exabytes of storage capacity will be required for human genomic data.

“Genomic data is really big data,” said Miller. “And it is really compute intensive. Processing a single human genome requires several hundred core hours. Computer processing units, on modern laptop computers typically have four cores. Data centers, by contrast, have hundreds of thousands of cores, which ‘make genomics a perfect cloud workload.’”

The Microsoft Genomics service is part of Healthcare NExT, a Microsoft initiative that aims to accelerate healthcare innovation through AI and cloud computing.