The Genome Is Clearest Where Diagnoses Overlap

01 — One child, several labels

A diagnosis describes the phenotype. It does not name the cause.

The same developmental disruption can appear as intellectual disability, autism, seizures, motor impairment or a combination. The overlap is clinically real, but it should not be read as one condition causing another.

02 — A like-for-like view

Intellectual disability carried the strongest sequencing signal

One meta-analysis applied the same review framework across 103 clinical sequencing studies. Yield was lowest in autism and highest in intellectual disability, but the wide intervals and high study heterogeneity warn against treating these as fixed probabilities.

Random-effects meta-analysis of 103 studies and 32,331 people: ASD 17.1% (95% CI 11–25), epilepsy 24.0% (22–27), ID 28.2% (22–35). Studies used gene panels or exome sequencing and were highly heterogeneous; these are cohort yields, not an individual forecast. Source: Stefanski et al., Epilepsia.

03 — The signal is in the overlap

When additional features appear, the chance of an answer rises

Three independent studies tell the same directional story. Each comparison stays inside its own study because the tests, recruitment and definitions differ.

Sources: Srivastava et al., Genetics in Medicine; Tammimies et al., JAMA; Moreno-De-Luca et al., JAMA. The three pairs use different tests and populations; compare direction within each card, not bar length across cards.

04 — The testing ladder

Higher resolution found more—but the yields are not additive

Karyotyping sees large chromosome changes. Microarray sees much smaller deletions and duplications. Exome sequencing reads protein-coding sequence. Each widened the diagnostic window, but each benchmark came from a different evidence base.

Do not add these percentages. Modern exome and genome pipelines can detect some copy-number changes, patients may receive several tests, and the recruited populations differ. Yield reflects both technology and who was selected for testing.

Cytogenetic benchmarks: Miller et al., American Journal of Human Genetics. Exome benchmark: Srivastava et al., Genetics in Medicine.

05 — More genome, a modest gain

Whole-genome sequencing consolidated the search. It did not solve most cases.

In 150 consecutive patient-parent trios, genome sequencing found every conclusive diagnosis produced by exome-based standard care—plus two more. Its advantage was completeness in one workflow, not a dramatic jump in answers.

Prospective parallel testing of 150 neurodevelopmental-disorder trios. Standard care integrated exome sequencing with other clinically indicated tests; WGS identified 45 conclusive diagnoses versus 43. Source: van der Sanden et al., European Journal of Human Genetics.

06 — A negative result can age

The DNA stays the same. The interpretation keeps moving.

A systematic review estimated that reanalysis produces an additional diagnosis in about one in ten previously negative exomes. New gene-disease links are often the reason.

The 10% estimate applies to previously unresolved cases with suspected Mendelian disorders, not to every negative test. A standard-care NDD reanalysis cohort reported a 13% yield. Source: van Slobbe et al., European Journal of Pediatrics.

07 — The interpretation frontier

Sequencing is no longer the only bottleneck

The genome is most informative when the phenotype points toward a rare, high-impact variant. It is less decisive for polygenic liability, acquired injury and variant classes that current workflows still struggle to interpret.

The frontier has shifted from reading DNA to knowing what a difference in DNA means—for this child, with this combination of features.

Variants of uncertain significance are not evenly distributed: genomic reference data remain substantially richer for people of European ancestry. Source: National Human Genome Research Institute.