Scaling Clinical Diagnostic Operations and Unlocking Genomic Innovation Using AWS with Natera

Natera specializes in cell-free DNA testing for oncology, women’s health, and organ health, providing insights to guide personalized care decisions. The company’s proprietary genomic datasets, in combination with real-world data, including clinical and demographic patient data, provide a comprehensive picture of patient journeys across various disease areas. This is useful to enhance precision medicine.

An area of priority for Natera was streamlining the extraction of insights from this data, a time-consuming process to get a cohesive view of what the team was looking for. To streamline this workflow and deliver more timely, actionable intelligence to guide decision-making, Natera aimed to create a centralized data and analytics solution within the company’s established compliance policies for defined user roles across the organization. “We wanted to shift from having to replicate data into other data solutions by giving users the opportunity to work on the data products in the data mesh,” says Mirko Buholzer, vice president of software engineering at Natera. “We were looking at AWS to form the backbone of the tooling and capabilities we needed.”

Natera used AWS Migration Acceleration Program (AWS MAP) to accelerate its cloud migration and modernization journey. As a part of this initiative, Natera transitioned a number of its variant calling pipelines to AWS HealthOmics, which helps transform genomic, transcriptomic, and other omics data into insights. The resulting genomic testing data is then stored in AWS HealthOmics variant store, and Amazon Athena, a simplified, flexible way to analyze petabytes of data where it lives, is used for querying it . Subsequently, this data is transformed in Amazon Redshift, which helps to modernize data analytics workloads and deliver insights for businesses. The processed genomic data gets integrated with real-world data that has been abstracted by Natera’s nursing teams.

To automate the extraction of data elements from clinical documents, Natera uses Amazon Bedrock—a fully managed service that offers a choice of high-performing foundation models—and Amazon Textract, which helps to automatically extract printed text, handwriting, layout elements, and data from any document. “Using Amazon Bedrock, we extract 46 data elements per disease type compared to 25 elements earlier,” says Alyssa Antonopoulos, senior director of data and informatics at Natera. “We chose Amazon Bedrock for its performance and for its security measures that can help us maintain HIPAA compliance.” Natera has also observed improved accuracy in data extracted through Amazon Bedrock, especially in tricky data elements such as medications and dates.

Once the data is extracted and integrated, the team uses Amazon QuickSight, which provides unified business intelligence, to surface insights to business users. In parallel, these products are published to a data mesh that Natera built using Amazon DataZone—which helps to unlock data across organizational boundaries with built-in governance capabilities—where they are accessible by multiple users. “Using Amazon DataZone, researchers have a single pane of glass to access data,” says Buholzer. “It brings everyone to one place to do their work.”

With this solution, Natera has demonstrated how real-world data can be seamlessly integrated into a complete data solution, using Amazon DataZone, AWS HealthOmics, and Amazon Bedrock. This provides for the use of diverse, high-quality datasets to drive more informed decision-making and unlock valuable insights that can lead to improved patient outcomes and accelerated test development.

The biggest impact comes from the integration of systems and centralization of data. Previously, it was difficult to have a coherent view of all available data, but users can now perform direct operations on all data that is on the centralized platform, removing the need to move data between different locations and improving overall efficiency. This centralization also makes for enhanced data governance and compliance, because Natera can monitor who subscribes to which data types.

Users have access to near real-time data insights and no longer depend on other teams for data access. “The pipeline is near real time,” says Antonopoulos. “We aren’t querying stagnant data anymore, and we can draw insights from the data much more rapidly.” Fluid access to data is unlocking new insights into Natera tests and their impact on patient journeys.

Faster innovation is another key win. “The innovation cycle went down from weeks and even months—depending on if the data wasn’t available and we had to collect or make it available—to days or hours,” says Buholzer. “We can do those queries in near real time now.”

Natera’s relationship with AWS is one of true, two-way collaboration. “It’s been great to work alongside AWS, and we appreciate their willingness to help us and meet our use cases,” says Buholzer. “The familiarity of our engineering teams with AWS also helped to scale this solution up quickly.”