Science1 distinct publisher3 min readUpdated
An Australian study flagged a fatal immune disorder in a baby before symptoms appeared. Dozens of pilots now face harder questions: cost, who interprets variants, and findings nobody can act on.
The Scientist · Science desk

Compiled by The ScientistSomething wrong?How this is made
Genomic newborn screening now has identified saves attached to named children, which moves the argument from whether sequencing finds anything to who runs it and who pays. Giselle Ghattas, now two, has familial haemophagocytic lymphohistiocytosis, a condition that causes fever and inflammation and can progress to organ failure, neurological damage and death in as little as months if untreated [1][2].
Her parents enrolled her in BabyScreen+, an Australian study that uses whole-genome sequencing to screen newborns for variants associated with severe, treatable diseases; they came across it on social media [4]. She received a bone-marrow transplant at six months old and, after some complications, has recovered, and according to her father her physicians expect nothing further beyond routine monitoring [5]. Because the disease is rare and variable, clinicians often misdiagnose it or fail to catch it early [3]. "If I had just kept scrolling on Facebook and not joined, then we'd probably still be, potentially even now, working out: 'What's wrong with her?'", her father Justin Ghattas told Nature [6]. That is the case for the technology and also the problem with the current delivery model: the save depended on a Facebook post.
The arithmetic explains why scaling is the hard part. Conventional screening uses a dried blood spot from the heel, analysed mostly by chemical assays of proteins and metabolites rather than by sequencing [9]. US guidelines recommend 66 conditions, mainly metabolic disorders; France screens for 16 and the United Kingdom for 10 [10]. Nearly 3.6 million infants are born in the United States each year, 98% are screened, and roughly 6,600, about 1 in 600, are predicted to test positive [11]. That is about 3.5 million screens a year and a positive rate near 0.18% [1][3]. Genomic pilots use DNA from the same blood spots and sequence hundreds of genes or whole genomes, with some panels covering more than 700 disorders [12], more than ten times the recommended US list [2].
Multiply an interpretation workload by that denominator and the constraints become financial and human rather than technical. Nature reports the process is currently costly and difficult to scale for broader implementation [15]. Robert Green, a medical geneticist at Harvard Medical School, says there is a lot of controversy around applying sequencing to thousands of people, including privacy and the potential for discrimination by insurance companies [16]. Wendy Chung of Boston Children's Hospital, a principal investigator on GUARDIAN, one of the largest genomic newborn-screening studies to date, frames it as adding a modality to a public-health programme that already leaves no one behind [13][14]. Both things can be true, and the second is the reason the first matters: universal programmes cannot quietly ration a step that costs more than the assay it supplements.
The reporting is also explicit that not every family has had the Ghattas family's experience [17], and the material supplied does not put a per-baby price on sequencing and interpretation, or name who is contracted to curate variants at population scale. Those are the numbers to demand before a health system commits.
Watch for GUARDIAN and BabyScreen+ to publish denominators rather than anecdotes: how many babies sequenced, how many findings returned, how many of those were treatable, and how many families were handed a risk with no intervention behind it. Watch whether jurisdictions running 10 to 16 condition panels [10] treat a 700-disorder genome as an expansion or a different programme with a different budget line. And watch the insurance and privacy rules, because Green's concern about discrimination is a policy gap, not a laboratory one [16].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Early results have shown that genomic screening approaches can flag treatable conditions that are not covered by conventional newborn screening, which checks for up to a few dozen conditions.
Not everyone has had the positive experience with genomic newborn screening that Giselle Ghattas's family has.
Giselle Ghattas, a two-year-old, has a rare genetic disorder known as familial haemophagocytic lymphohistiocytosis (HLH).
HLH causes fever and inflammation and can spiral into organ failure, neurological damage and death in as little as months if it goes untreated.
Because HLH is rare and variable, clinicians often misdiagnose it or fail to catch it early.
Giselle's parents, Justin Ghattas and Scarlett Morwood, enrolled her in BabyScreen+, a study in Australia that uses whole-genome sequencing to screen newborns for genetic variants associated with severe, treatable diseases; they came across the study on social media.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Named investigators and quantified pilot data, but one publisher and preliminary results
The single supplied source is a science-press feature that names institutions and principal investigators (Chung/GUARDIAN, Green/BabySeq) and carries specific, referenced figures: panel sizes of 66/16/10 conditions, 700+ disorder genomic panels, 98% conventional coverage of ~3.6M US births, and cohort-level yields. That is substantially better than assertion-only reporting. It is capped well below high confidence because the headline GUARDIAN numbers are explicitly preliminary conference data on 15,000 of a planned 100,000 participants, the clinical win is a single named case, and no second publisher in the cluster corroborates anything.
Pilot-stage across dozens of programmes, sitting beside near-universal conventional screening
Genomic newborn screening exists at research-cohort scale: dozens of feasibility initiatives, GUARDIAN at 15,000 of a planned 100,000, BabySeq at ~1,045 infants over more than a decade, and BabyScreen+ producing individual case saves. That is real, multi-site, patient-affecting deployment rather than a demo, which lifts the score above minimal. It stays low because none of it is routine practice: the installed base is still the conventional protein/metabolite panel that reaches 98% of US newborns, and the source names cost and scale-up difficulty as unresolved.
Mildly overstated: anecdote-led framing ahead of preliminary, pilot-scale data
The narrative leads with a single thriving toddler and the prospect that sequencing could 'revolutionize' newborn screening, while the underlying evidence is a 2.7% confirmed-finding rate from preliminary data on 15,000 of 100,000 planned participants and an ~11% variant rate in 432 sequenced BabySeq infants. That gap is small rather than large because the same source discloses the counterweights itself: cost and scale-up problems, results-not-diagnoses workflows needing confirmatory testing, privacy and insurer-discrimination controversy, and an explicit note that not all families have had a positive experience. Positive but modest.
Principal investigators of the programmes are the main voices
The two expert voices are directly invested in the field they assess: Wendy Chung is a principal investigator on GUARDIAN and supplies the 'enhance a successful public-health initiative' framing, and Robert Green co-led BabySeq. Programme leaders describing their own programmes' promise is a real interest to disclose. The score is held near mid-range because the source labels both affiliations plainly, and Green is quoted raising controversy against his own field's expansion rather than only promoting it; no funding, vendor or commercial-sequencing relationships are disclosed either way in the supplied excerpt.
Moderate: credible specifics from a single publisher on a pre-implementation field
Confidence is moderate. The factual spine - screening mechanics, national panel sizes, cohort sizes and yields, named barriers - is specific and internally consistent, and the article is transparent about the preliminary status of its strongest numbers. But the cluster has one publisher, so there is no independent corroboration; the key clinical outcome is one case; the negative experiences alluded to are not detailed in the supplied text; and cost, workforce and false-positive burden are unquantified, which limits how far any forward-looking read can be pushed.
product
Fatal crashes rise 15% on big album release days. That is a media-selection design problem.1 distinct publisher
product
Aircrew outrank nuclear technologists in radiation-linked cancer deaths, JAMA study finds1 distinct publisher
science
Caltech model says fast-dividing tissue may be proofreading, not just accumulating risk1 distinct publisher
science
The Pentagon wants to test soldiers' testosterone. The label for treating it is narrow.1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 18, 2026