Sonrai Accelerates Single-Cell RNA-seq Data Analysis Using Amazon Bedrock

As a pioneer in biotech offerings, Sonrai provides advanced analytics solutions tailored to the needs of its clients in the biotech and pharmaceutical industries. The company specializes in diverse cloud technologies for precision medicine, a complex approach that considers a patient’s unique genes and environment in the treatment of disease. Its flagship product, Sonrai Discovery, empowers researchers to analyze and interpret complex biological data, such as scRNA-seq technology that lets researchers study gene expression in individual cells.

Traditionally, the annotation of cell clusters in scRNA-seq datasets is a manual, time-consuming task that requires specialized knowledge from immunologists. This process can delay critical research and decision-making in drug discovery and precision medicine. Sonrai identified the need for an automated solution that could perform these annotations quickly and accurately, reducing reliance on domain experts and speeding up research timelines. Additionally, the company recognized that any solution needed built-in stringent data governance and security measures. “We needed to find a way to empower researchers to gain insights faster without being bogged down by the labor-intensive process of data annotation,” says Dr. Matthew Alderdice, head of data science at Sonrai. “Our goal was to help them focus on discovery and innovation rather than manual data annotation.”

Sonrai built its technology stack for scRNA-seq analysis entirely on AWS, resulting in faster analysis that supports the development of more targeted therapies to ultimately help patients. Using Amazon Bedrock to power Sonrai Discovery, Sonrai has transformed its approach to cluster annotation. Rather than relying on researchers to select and label cells by hand, Sonrai Discovery uses large language models through Amazon Bedrock, reducing errors fivefold while improving consistency and speed. “Using Amazon Bedrock, we have been able to cut annotation times by up to 50 percent, freeing up our clients to focus on what matters most: scientific discovery,” says William Guesdon, senior bioinformatics scientist at Sonrai. In addition to significant reduction in annotation times, domain experts benefit from the automated generation of text reports, which simplifies the review process. Plus, Sonrai saves its clients up to 20,000 dollars per experiment, based on the average salary of an expert immunologist.

As part of its solution, Sonrai needs to handle and store Fast Quality (FASTQ) files, raw sequence data, and quality scores that result from the high-throughput scRNA-seq analysis. For FASTQ file processing, Sonrai uses AWS HealthOmics, a purpose-built service that helps healthcare and life science organizations and their software partners store, query, and analyze genomic, transcriptomic, and other omics data and then generate insights from that data to improve health.

The company also uses AWS to prepare its datasets for Sonrai Discovery. Sonrai handles data processing and modeling through the use of Amazon SageMaker, which provides fully managed infrastructure, tools, and workflows to build, train, and deploy machine learning models for any use case.

Sonrai defines and deploys its infrastructure automatically using AWS Cloud Development Kit (AWS CDK), which lets organizations define their cloud application resources using familiar programming languages. For scalable and cost-effective data storage, Sonrai uses Amazon Simple Storage Service (Amazon S3), an object storage service offering scalability, data availability, security, and performance. All told, Sonrai’s technology stack on AWS has enhanced its capabilities, driving faster insights that enrich precision medicine.