Amazon Web Services and Fabric Genomics Case Study Highlights Performance and Scalability

Fabric Genomics is a software genomics company that provides end-to-end genomic data analysis, annotation, curation, classification, and reporting solutions to clinical labs, hospital labs, life science companies, country sequencing programs, and research institutions. The organization’s machine-learning technologies and advanced proprietary algorithms VAAST and Phevor enable scientists to quickly and accurately process and analyze whole genomes, exomes, and gene panels for hereditary disease and oncology. Headquartered in Oakland, California, Fabric Genomics helps its customers accelerate access to insights related to the cause of genetic diseases.

“Using Fabric Genomics’ interpretation software on AWS, along with other software tools, we can interpret an entire genome’s variant set within minutes, whereas other platforms might take hours,“ said Shareef Nahas, Ph.D., Senior Director of Clinical Laboratory Operations, Rady Children’s Institute for Genomic Medicine. “That means we can start our analysis quickly and distill millions of data points into a few that are distinct markers for diseases in children.”

The Challenge

Each day, customers using Fabric Genomics’ platform have the potential to help save lives by identifying which variants are disease-causing. This information helps clinicians to determine the best treatment and medications for their patients based on the patients’ genetics.

Strong performance is key to the platform’s success. “Speed is so important for our software because our customers are working on pinpointing the genetic causes of disease in very ill patients,” says Martin Reese, chief executive officer and co-founder of Fabric Genomics. For example, Rady Children’s Institute for Genomic Medicine uses Fabric Genomics interpretation and clinical reporting software as part of its bioinformatics pipeline for identifying disease-causing variants in a child’s genome. “We perform rapid genome sequencing to find genetic disorders that cause critically ill kids to be admitted to neonatal or pediatric intensive care units,” says Shareef Nahas, the institute’s senior director of clinical laboratory operations. “The faster we can sequence the genomes and deliver the data to medical staff, the sooner these children can be treated.”

In addition to performance, Fabric Genomics needs its software to be highly scalable. “Scalability is very important for us, especially considering that clinical labs often need to sequence the DNA of a patient’s parents and siblings,” says Reese. “The data files are massive—millions of files up to 200 gigabytes each—and they need to be compared very quickly, so we need tools that support large files and the ability to analyze them all at once.”

Why Amazon Web Services

Fabric Genomics chose to run its software on the Amazon Web Services (AWS) Cloud to gain the performance and scalability that it needs. “By using AWS, we get the computational power and capacity we require to process and analyze large genomic datasets,” Reese says. “In a typical laboratory today, it takes 12 weeks to interpret a genome. With our software, we have reduced that to two hours.”

The company runs two computational workloads—variant discovery and variant interpretation—on AWS, using Amazon Elastic Compute Cloud (Amazon EC2) instances, which run on the latest Intel® Xeon® processors. “The Intel processors are critical to the performance of our workloads,” says Reese. “In some of our more complex variant discovery processes, we use up to 200 gigabytes of RAM in all processing cores simultaneously, and the Intel Xeons perform very well under that kind of strenuous load.” Fabric Genomics stores genomic sequencing results in Amazon Simple Storage Service (Amazon S3) buckets, and it uses high-throughput Amazon Elastic Block Store (Amazon EBS) volumes for a reference database.

After genomic data is sequenced, a FASTQ or VCF file is uploaded into Fabric Genomics’ platform for analysis of the variants. Using machine-learning technology and advanced algorithms, variants are ranked. Lab directors can review the data and generate a customizable clinical report.

Fabric Genomics’ platform delivers comprehensive data and network security and full support for compliance with regulations such as HIPAA, ISO 27001, and FDA 21 CFR Part 11. In addition, the organization relies heavily on the security capabilities of AWS – as well as additional encryption technologies – to protect patient data. “This is some of the most sensitive data in the world—it’s literally the code of who a person is,” Reese says. “AWS gives us enhanced security to help us comply with strict patient health information (PHI) data-protection requirements.”

The Benefits

AWS and Intel technologies give Fabric Genomics the ability to offer fast computational speeds and security to its customers. “Our software can perform genetic sequencing and analysis very quickly because of the performance we get with AWS and Intel,” says Reese. As a result, customers like Rady Children’s Institute of Genomic Medicine are getting data to physicians faster. “Using Fabric Genomics interpretation software on AWS, along with other software tools, we can interpret an entire genome’s variant set within minutes, whereas other platforms might take hours,” says Nahas. “That means we can start our analysis quickly and distill millions of data points into a few that are distinct markers for diseases in children,” says Nahas. “Because of this, we can sometimes find a diagnosis within minutes and get the information to the medical staff at Rady Children’s Hospital right away, so they can treat critically ill children.”

In some cases, medical teams at Rady Children’s Hospital using the Rady Children’s Institute for Genomic Medicine’s bioinformatics pipeline for rapid whole genome sequencing have discovered that a patient has a different disease than originally thought. “We can make more accurate diagnoses because our analysis has shown what the disease is, and physicians can know the exact medications or treatments that could help,” says Nahas.

Fabric Genomics is taking advantage of the scalability of AWS to respond to customers with very large data-analysis needs. “We can scale our platform as widely as we want to by using AWS,” Reese says. “If a customer wanted to process 10,000 genomes, we could process all those genomes simultaneously. We have an unlimited ability to expand compute capacity. Without the scalability of AWS, we wouldn’t be able to deliver 10,000 reports in two days.” In addition, the company is using multiple AWS Availability Zones worldwide to meet its customers’ needs. “AWS is in so many regions, we can deploy our software globally whenever we want to,” Reese says. “That helps us because some of our customers have sensitive data that cannot leave their region due to local regulatory requirements. We can help them store that data in their country or region by using AWS.”

Fabric Genomics’ platform is being utilized by clinical labs for analysis of whole genomes, exomes, and panels, and the company is excited to partner with many labs as they expand their testing volume to NGS testing. “Precision medicine is just at the beginning right now,” says Reese. “The more people use AWS, the easier it will be to share data, so people can find new insights and we can treat patients earlier and identify the cause of disease. That will allow us to be more effective with our treatment options, and therefore have healthier patients going forward.”

Learn More

Read the AWS Fabric Genomics case study to learn more.

Date: March 26, 2018