San Diego – Dec. 5, 2024 – Two studies published today in the American Journal of Human Genetics show the potential for genomic screening in newborns to address high rates of infant hospitalization and mortality in the United States. Presently, hundreds of genetic diseases are either preventable or treatable but currently are detected only after a child falls ill and endures a years-long “diagnostic odyssey,” often receiving diagnoses too late to achieve the best outcomes.
The first study, titled “Prequalification of genome-based newborn screening for severe childhood genetic diseases through federated training based on hyperselection” described a novel platform with scalability and performance that will allow millions of babies to be screened and treated by genome sequencing and artificial intelligence within two weeks from birth.
Previously, medical genome sequencing was much too expensive for newborn screening. Plus, a high rate of false positive results – genome findings that falsely suggest a newborn to have a genetic disease – have been of great concern when genomes are examined without clinical context, Previously, no method existed to translate genome results into treatment guidance in a way that most physicians could understand and put into practice.
The novel platform, BeginNGS (pronounced beginnings) solves this. It makes use of the latest genome sequencing technology to provide an affordable genome. BeginNGS uses a combination of human and artificial intelligence tools to automate the complex process of interpreting disease risk based on genome information alone, which is critical for scaling to the 3.7 million babies born in the U.S. each year.
The study reported a 97 percent reduction in false positives based on a method derived from human evolution. The genome variations that cause severe childhood diseases are subject to extreme natural selection called purifying hyperselection. As a result, DNA variants that truly cause severe childhood disease are not found in genomes of elderly persons. By studying the genomes of almost half a million middle aged and elderly subjects, from the UK Biobank and Mexico City Prospective Study, researchers were able to discover those false positive DNA differences and reduce their occurrence to less than 1 in 50 subjects tested.
The computational methodology uses query federation, a method to analyze genomes remotely without data being moved or shared, which is enabled by TileDB, a database technology partner for BeginNGS. Remarkably, after removing these DNA variants, BeginNGS retained greater than 99 percent sensitivity when compared with the gold standard method of rapid diagnostic genome sequencing.
BeginNGS used a custom-built clinical guidance system called Genome to Treatment (GTRx) to communicate a potential course of action for babies who screen positive. Many of these disorders are so rare the typical physician will rarely see them in practice. GTRx provides practical guidance for physicians in a manner that is easy to understand. Testing of over 3,000 children with suspected genetic diseases revealed that 1 in 14 would have benefited from BeginNGS by receiving a time-to-diagnosis of 121 days earlier than compared with gold standard testing after those children developed symptoms Testing of over 3,000 children with suspected genetic diseases revealed that 1 in 14 would have benefited from BeginNGS and would have received diagnoses and treatment 121 days earlier than the current gold standard, which is to test after symptom onset. In addition, testing revealed that BeginNGS would have benefited 1 in 13 babies who died in infancy.
"The future of newborn genetic screening lies in global collaboration and shared data resources," says Stavros Papadopoulos, CEO and Founder of TileDB. "By connecting genetic information across international databases, we significantly enhance our ability to identify and understand rare diseases — an endeavor that transcends individual projects and geographical boundaries. Through TileDB's expansion of the BeginNGS consortium and our federated query capabilities, we're enabling more comprehensive analysis of variant datasets. For RCIGM and the families they serve, this translates directly into faster, more reliable answers during those critical early days of life."
The second study, entitled “Genome-based newborn screening for severe childhood diseases has high positive predictive value and sensitivity in a NICU pilot trial” evaluated whether BeginNGS was ready for broader expansion. In this trail, 120 babies in the neonatal intensive care unit at Rady Children’s Hospital - San Diego, received the BeginNGS screening. Results were compared with traditional, federally mandated newborn screening and t\rapid diagnostic genome sequencing which evaluated all ~10,000 genetic diseases.
“The amazing, unexpected result of this BeginNGS trial was that nearly 30 percent of NICU babies who weren’t considered to need genome sequencing actually had genetic diseases -- this is similar to the rate of diagnosis in babies who are suspected of having genetic diseases! said Stephen F. Kingsmore, M.D., DSc, president and CEO of Rady Children’s Institute for Genomic Medicine “This suggests that the health benefits of rapid whole genome sequencing apply to every baby admitted to a Level IV NICU, not just those who are currently being tested.”
Only babies who were not suspected of having genetic diseases were eligible for enrollment in the clinical trial since the trial wished to mimic screening of healthy newborns. BeginNGS genome-based newborn screening was shown to be safe and effective. One in 24 babies tested had positive results that were likely to impact their care. BeginNGS had no false positives, showing that the purifying hyperselection methods indeed worked in the real world. Eighty four percent of parents in the trial reported that their child’s genomic sequencing results were useful, and 80 percent felt that participation did their child a lot of good. When compared with state newborn screening, BeginNGS had a higher true positive rate and lower false positive rate.
“Genome-based newborn screening has the potential to transform health outcomes for children with certain rare diseases by accelerating their time to diagnosis and proper care,” said Tom DeFay, PhD, Vice Chair of BeginNGS and Deputy Head of Diagnostics at Alexion, AstraZeneca Rare Disease. “As a founding member of the BeginNGS Consortium, Alexion is encouraged by these Phase 2 results and remains committed to advancing health equity by helping improve diagnostics for families impacted by rare genetic and often life-threatening conditions.”BeginNGS technology partner, Illumina, provided the DRAGEN secondary analysis pipeline and bioinformatics expertise to process data for the studies. Illumina’s sequencing instruments and DRAGEN analysis provide highly accurate baseline data on which to build a gNGS-based NBS program with demonstrated potential to identify infants with treatable genetic conditions.
These studies pave the way for a much larger, multicenter clinical trial to formally compare BeginNGS with standard newborn screening. That trial has started and to date is replicating the findings of the pilot study. In addition, now that computational methods exist for removal of false positives, BeginNGS is poised to expand from 412 severe childhood genetic diseases to the more than 2,000 disorders that have been suggested to be actionable early onset rare disorders for newborn screening. Now that the feasibility of federated queries has been successfully demonstrated it will be possible to expand these to many genome biobanks worldwide to examine the incidence and prevalence of genetic diseases across the globe, allowing BeginNGS be tailored to screen each population.
###
About Rady Children’s Institute for Genomic Medicine:
We are transforming pediatric critical care by advancing personalized medicine in healthcare for infants and children with rare diseases. The Institute uses the power of genomic data coupled with unprecedented speed and accuracy to enable the rapid diagnosis and targeted treatment of critically ill patients at Rady Children’s Hospital-San Diego and a growing network of more than 100 health systems nationwide. The vision is to expand access of genomic medicine to improve outcomes for children. RCIGM is a non-profit research institute embedded within Rady Children’s Hospital and Health Center. Learn more at RadyGenomics.org. Follow us on LinkedIn.
Editors Notes and Definitions:
Query: In the context of a database, a query is a request for information or data made by a user and written in a specific format.
BeginNGS Query: A long list of putatively disease-causing variants.
Result Set: A short list of observed diplotypes and counts.
The output is a text file with the variant info and metadata from the BeginNGS query.
Data Federation: A data integration technique that provides a unified view of data from disparate sources without requiring physical data movement or consolidation.
Federated Learning (also known as a form of "collaborative learning"): A machine learning technique focusing on multiple groups collaboratively training a model while ensuring while ensuring data remains decentralized and has data heterogeneity.
Decentralized refers to a model where data is stored, processed, and managed across multiple locations without a central server or authority controlling all data flows.
Data heterogeneity (i.e. the data has variability and is dissimilar).
END
