Scaling newborn genomic screening using AWS with Regione Puglia Sanità

Puglia, a region of nearly 4 million people in southern Italy, records roughly 24,000 to 25,000 births each year. Before Genoma Puglia, families affected by rare genetic diseases could enter a years-long journey involving repeated tests, specialist visits, and prolonged uncertainty, with narrowing opportunities for early intervention. “In rare genetic disease, the long diagnostic journey is called the diagnostic odyssey,” says Simone Gardini, CEO and cofounder at GenomeUp. “The key point is reducing the time it takes.” Puglia had already introduced a screening program for spinal muscular atrophy (SMA) in 2021, and a 2023 regional law supported the launch of a genomic screening pilot. The pilot screened more than 4,000 newborns and demonstrated the potential to extend genomic screening across the region to all newborns.

Moving from a pilot to a region-wide program created a challenge. Regione Puglia Sanità needed to connect 24 maternity wards across a long, narrow region to a single laboratory workflow. It also meant keeping sample identity secure and traceable from the moment of collection through final report. The process also had to support hundreds of conditions and a high weekly sample volume without requiring repeat blood draws. Because participation was optional, the program needed to earn families’ trust while giving clinicians a dependable way to register samples, review findings, and prioritize cases requiring rapid attention.

Regione Puglia Sanità worked alongside GenomeUp because of the Partner’s ability to customize an entire genomics workflow and collaborate closely with a hospital’s genetics and bioinformatics specialists. Over approximately 1 year, the teams adapted JuliaOmix to support each stage of the program. The system is organized into three connected modules: JuliaOmix Track (JOTRK) manages sample check-in, work-plan creation, and traceability; JuliaOmix Cloud (JOCLD) covers run monitoring, primary analysis, and storage of sequencing data; and JuliaOmix Laboratory (JOLAB) performs secondary and tertiary analysis, variant prioritization, and report generation. Each dried blood spot sample is identified only by barcode and pseudonymized throughout the workflow.

JuliaOmix runs on AWS. GenomeUp uses AWS Identity and Access Management (AWS IAM) to securely manage identities and access to AWS services and resources alongside AWS Secrets Manager to centrally manage the life cycle of secrets used across the pipeline. The platform stores raw sequencing files and analysis outputs in Amazon Simple Storage Service (Amazon S3), which offers industry-leading scalability, data availability, security, and performance to support the terabytes of genomic data that are generated each week. It uses Amazon S3 multipart upload to upload large objects as a set of parts, with parallel processing improving throughput. Sequencing and variant-analysis workloads run on AWS Lambda for serverless compute and AWS Batch for batch processing for machine learning model training, simulation, and analysis. The platform uses Amazon DynamoDB—a serverless, fully managed, distributed NoSQL database—to calculate and store the local frequencies of genetic variants that are identified by individual genetics laboratories, as well as aggregated data across the entire organization. Meanwhile, Amazon Aurora serverless—an on-demand, autoscaling configuration for Amazon Aurora, a relational database service—stores patient and sample information and provides complete traceability for tracking the extraction of genetic material from biological samples and the subsequent sequencing processes. “AWS provides the infrastructure for our system and software, helping us focus on JuliaOmix while continuing to build on the AWS stack,” says Nicolò Risitano, chief technology officer at GenomeUp.

The clinical and GenomeUp teams created a three-level variant-analysis process. The first level flags samples with known pathogenic variants. The second reviews all samples for pathogenic and likely pathogenic variants. The third examines genes associated with recessive disorders. This structure helps clinicians prioritize findings requiring the fastest attention while continuing the review of every sample. The laboratory analyzes approximately 400–450 samples each week and can pool up to 384 samples into a single sequencing run.

From the region-wide program launched in April 2025 through the end of June 2026, Genoma Puglia screened 30,214 newborns. Participation was 90.6 percent, and the program gives Genoma Puglia the ability to extend testing to every newborn in the region at a cost of just €390 per child. “We have 1.8 percent sequencing repetition, and in no cases has it been necessary to perform another dried blood spot analysis on a newborn,” says Mattia Gentile, head of genetics at Di Venere Hospital in Bari, Italy, the hospital coordinating and operating the program. To date, the program has returned 742 positive cases, with the reported variants confirmed through follow-up testing. Despite initially relying on only a phone number collected at birth, the clinical team has also contacted and managed families in over 98 percent of positive cases.

By screening 433 genes for more than 500 genetic conditions for every newborn, the program expands the range of diseases that can be identified before symptoms appear. The clinical team reports that early findings have supported earlier treatment, preventive care, and clinical follow-up. Genomic screening has also helped resolve the exact diagnoses or exclude the disorder for more than 100 cases that were positive in conventional metabolic newborn screening rapidly. This means that clinicians can stop unneeded examinations, follow-up, or treatment and instead prioritize those families that require rapid follow-up. As Genoma Puglia continues to expand, Regione Puglia Sanità is helping more families move from years of uncertainty toward more timely care.

GenomeUp is an Italian bioinformatics company founded in 2017 and the developer of the JuliaOmix genomics solution.