Genome sequencing in developmental delay: what a new trial found

Imagine a seven-year-old boy. He is autistic, has a moderate learning disability, and had a normal microarray a few years ago, when that was still the standard first test. Both parents are well.

In 30 years as a doctor I have met many families in that position, so a new Belgian trial caught my eye. It asked whether reading the entire genome finds more answers.

Think of DNA as a vast cookbook in which the recipes for proteins fill only a small fraction of the pages. Exome sequencing reads just the recipes. Genome sequencing reads the whole book, margins and all. That ought to find more.

All eight genetics centres in Belgium took part. Between 2021 and 2024 they recruited 567 people with unexplained developmental disorders, mostly moderate to profound intellectual disability, or milder disability with unusual physical features or congenital anomalies. About one in six were adults. Nobody had already had an exome or a large gene panel.

Participants were randomly assigned. Half had what counted as standard care in Belgium: a check for missing or extra chunks of chromosome (microarray or a low-coverage genome), followed by an exome. The other half had whole genome sequencing. Both parents were tested wherever possible.

The genome came out ahead. A definite diagnosis was found in 39.8% of the genome group (113 of 284), against 30% with standard care (85 of 283). That 9.8-point gap was statistically significant.

Females were diagnosed more often than males (45.5% against 28.5%), and the standard arm happened to contain more males. The genome arm also had extra software reanalysis of negative cases, and six of its diagnoses came from wider gene lists, looser filtering and newer literature rather than from the genome itself. Allow for both and the gap shrinks to 7.3 percentage points, no longer statistically significant (p = 0.069, so the difference could be due to chance). The trial recruited 567 of a planned 800, limiting its power. The same laboratories interpreted both arms, and the paper describes no blinding.

Most of the extra diagnoses were not in the margins. They were ordinary typos in the recipes, especially small insertions and deletions, at positions the exome had read perfectly well. Part of the difference seems to come from how the data were processed and interpreted.

The margins, the parts of the genome outside the recipes, gave three extra findings: one case of fragile X syndrome, caused by a repeated stretch of DNA, a change just upstream of the MEF2C gene and, on later reanalysis, a diagnosis in RNU4-2, a non-coding gene only linked to developmental disorders in 2024.

Autism was not an entry criterion, but about a third of participants had autism or autistic behaviour recorded. In that group roughly one in four (24.1%) received a diagnosis. All had intellectual disability or congenital anomalies, so the trial says nothing about autism alone.

Now back to that boy with two well parents. With both parents tested, laboratories usually hunt for brand-new changes that neither parent carries, X-linked changes, and recessive conditions where both parents are silent carriers. This team also looked for changes in dominant genes handed down from a parent.

They found 22. These added 3.9 percentage points to the overall yield and made up 11% of all definite diagnoses. In 18 of the 22 families nobody had suspected a dominant condition.

A parent can carry the cause and still be well. Half the carrier parents (11) had no symptoms, eight had mild features and three had more significant ones. Two were probably mosaic, carrying it in only some cells. In every gene involved, the same change can cause severe problems in one person, mild ones in another and none at all in a third.

Two mothers turned out to carry PTEN variants. PTEN changes cause Cowden syndrome, with a raised cancer risk and a need for lifelong surveillance. A test ordered for a child ended up bearing on each mother's own health. A dominant change in a parent usually gives a 50% chance of passing it to each future child.

One more caveat: Illumina, which makes the sequencers, supplied the genome-arm consumables and its Emedgene software, and an Illumina employee reviewed the manuscript.

In England, the separate intellectual disability tests (R29 and R377) left the NHS National Genomic Test Directory on 1 April 2026. Whole genome sequencing is now the standard first-line test for these children and goes through R27 Paediatric disorders, which community paediatricians can request. Eligibility covers unexplained moderate, severe or profound developmental delay or intellectual disability, or congenital malformations and/or dysmorphic features. Autism alone, and non-syndromic mild delay or intellectual disability, do not qualify. So in England, a child with moderate or greater delay, or with congenital anomalies and/or dysmorphic features, who has never had a genome may now be eligible.

Genome data are stored, so as new genes such as RNU4-2 are discovered, earlier results can be reanalysed without another blood sample.

In Germany, genome sequencing for rare diseases and cancer is offered through a national pilot, the Modellvorhaben Genomsequenzierung, set up under section 64e of the statutory health insurance code (SGB V). It has run since July 2024 and by law lasts at least five years. Access is through the centres for rare diseases (Zentren für Seltene Erkrankungen), whose case boards recommend who is sequenced.

Germany has no fixed list like R27. Sequencing needs a suspected rare disease where it is expected to add substantially to the diagnosis, confirmed by a multidisciplinary case conference at a participating hospital. Parents can be sequenced too if they consent. The German genetic testing law (Gendiagnostikgesetz) applies, so a doctor must explain the test and get written consent first. If a result points to a change inherited from a parent, that parent may need advice of their own.

Families who think their child may qualify can discuss testing with their paediatrician or local clinical genetics service. This post is general information, not advice about any individual child.

References

1. Geysens M, Souche E, Baroni MC, et al. A nationwide prospective randomized trial for diagnosing developmental disorders demonstrates genome sequencing outperforms standard of care. Genome Medicine. 2026;18:144. doi:10.1186/s13073-026-01771-2
2. Leeds Teaching Hospitals / North East and Yorkshire Genomic Laboratory Hub. Important: changes to National Test Directory from April 2026. https://www.leedsth.nhs.uk/wp-content/uploads/2026/03/Changes-to-microarray-and-WGS-National-Test-Directory-from-1st-April-2026.pdf
3. East Genomics. Rare disease directory update (April 2026). https://www.eastgenomics.nhs.uk/for-healthcare-professionals/updates/rare-disease-directory-update-april-2026/
4. NHS England. National genomic test directories. https://www.england.nhs.uk/publication/national-genomic-test-directories/
5. § 64e SGB V: Modellvorhaben zur umfassenden Diagnostik und Therapiefindung mittels Genomsequenzierung bei seltenen und bei onkologischen Erkrankungen. https://www.gesetze-im-internet.de/sgb_5/__64e.html
6. GKV-Spitzenverband. Modellvorhaben Genomsequenzierung nach § 64e SGB V. https://www.gkv-spitzenverband.de/krankenversicherung/forschung_modellvorhaben/mv_genomsequenzierung/genomsequenzierung.jsp

Dr Odet Aszkenasy is a Consultant Community Paediatrician and the author of The Genetics of Autism: A Guide for Parents and Professionals.