89 genes survived scrutiny. 426 didn’t.
A review of 515 genes linked to cerebral palsy found only 89 with strong evidence supporting the association.
By The Weekend · · 5 min read
Cerebral palsy affects up to three in every 1,000 babies born worldwide, making it the most common cause of childhood physical disability. A new analysis suggests much of what doctors think they know about its genetic causes doesn't hold up: of 515 genes previously reported as linked to the condition, only 89 have statistical evidence strong enough to support that claim.
515 genes went in. 89 came out the other side.
Researchers at the Jackson Laboratory for Genomic Medicine, led by computational biologist Peter Robinson and biocuration researcher Adam Arterbery, reviewed 21 published genetic studies of cerebral palsy. Between them, those studies had named 515 genes as potentially causal. The team then checked each gene against more than 5,500 other published studies, using statistical methods designed to separate genes with real disease associations from genes that turned up incidentally — present in a study's dataset, but not actually driving the condition.
The gap was not small. Only 89 of the 515 genes cleared the bar. That's 17 percent of the candidates the field has been treating, in various papers and clinical discussions, as plausible genetic causes of cerebral palsy.
The study, published in the American Journal of Human Genetics, also included a second test. The researchers ran whole-genome sequencing on 460 children diagnosed with cerebral palsy and found pathogenic or likely pathogenic gene variants in about 16 percent of them — variants spread across 60 different genes. When they checked those 60 genes against their own literature analysis, only 16 had statistical evidence linking them to cerebral palsy specifically.
Why the shortfall is bigger than the headline number suggests
A drop from 515 to 89 sounds like a rounding error correction. It isn't, because of what the smaller number implies about clinical practice already in motion. Genetic testing for cerebral palsy has been marketed, in part, on the promise that a positive result explains the condition — gives a family a cause, a name, sometimes a prognosis. If a child's test comes back with a variant in one of the 426 genes that didn't survive this analysis, the honest answer is that nobody currently has strong evidence that variant causes cerebral palsy at all. It might. It might also be an incidental finding — present, but irrelevant to the symptoms in front of the clinician.
The second-phase result sharpens this. Sequencing found disease-linked variants in 16 percent of the 460 children studied — a number that sounds like a genetic explanation was found for one in six cases. But only 16 of the 60 implicated genes had solid backing. So the real, defensible figure — variants in a gene with a demonstrated cerebral palsy association — is smaller than the headline 16 percent, though the researchers haven't published the exact recalculated share.
Robinson and Arterbery told ScienceAlert they expected some genes to have stronger evidence than others going in. What surprised them was the scale of the discrepancy — not a modest thinning of a large list, but a list reduced by over 80 percent.
The strongest pushback: absence of proof isn't proof of absence
The researchers are explicit that their analysis does not clear the 426 excluded genes of involvement. A gene can have a real, causal role and still lack "sufficient statistical evidence" in a review built from published literature — because the studies reporting it were small, because the variant is rare, or because nobody has yet run the specific analysis that would surface a stronger signal. Absence of statistical support, in a field this young, is a measurement gap as much as a verdict.
That caveat matters because it changes what the study is actually arguing. It is not saying cerebral palsy has no genetic basis, or that genetic testing is worthless. It's making a narrower claim: that the current evidence base linking many specific genes to the condition is thinner than the clinical and research conversation has treated it as being. The distinction between "we found no strong evidence" and "we found evidence of no link" is the whole difference between caution and dismissal — and the researchers land firmly on the cautious side.
The reframe that follows: a feature, not a diagnosis
The paper's more consequential proposal isn't the gene count — it's a suggested change in how cerebral palsy should be classified. The researchers argue it may be better understood "as a feature of certain Mendelian diseases rather than a separate disease" — meaning cerebral palsy's movement symptoms might, in many cases, be one component of a broader genetic condition rather than a standalone diagnosis with its own genetic cause.
That reframing has a practical consequence for how children get tested and treated. If cerebral palsy is a shared symptom across multiple underlying diseases rather than a single genetic entity, the useful question for a clinician shifts from "which cerebral palsy gene does this child have" to "which broader condition is producing this presentation, and what else does that condition predict." The second question opens doors the first one can't — connecting a movement disorder to other symptoms that might emerge later, or to treatments developed for the underlying Mendelian disease rather than for cerebral palsy as a category.
None of this changes the near-term reality for the roughly one in six children in the 460-person cohort who received a genetic result during their diagnostic workup. Families in that position are the ones who will feel the gap between "a variant was found" and "a variant with 89-gene-strength evidence was found" most directly, in the conversations genetic counselors have after a sequencing report lands. The study doesn't resolve those conversations. It gives them a clearer number to be honest about.