Mechanisms linking Parkinson’s disease genetic risk alleles to neurodegeneration

Determine the causal mechanisms by which Parkinson’s disease genetic risk alleles identified in population studies, including genome-wide association studies that implicate hundreds of loci of modest effect, lead to the neurodegenerative processes characteristic of Parkinson’s disease.

Background

Parkinson’s disease (PD) includes a small proportion of familial cases explained by high-penetrance variants, but most genetic risk arises from numerous common alleles of modest effect identified by genome-wide association studies. Despite extensive genetic findings, the field lacks a clear understanding of how these risk alleles produce the neurodegenerative phenotypes seen in PD.

The paper investigates whether a graph-based AI model (Proton) can infer links between PD genetic risk loci and molecular mechanisms, including those operating in dopaminergic neurons. This addresses the broader unresolved question of translating statistical genetic associations into mechanistic pathways driving PD pathology.

References

Only 5-10% of people with PD have familial forms explained by high-penetrance causal variants (SNCA, LRRK2, VPS35, PRKN, PINK1, DJ1), whereas genome-wide association studies (GWAS) implicate hundreds of loci of modest effect. How these risk alleles give rise to the neurodegenerative processes observed in PD remains unclear.

— Graph AI generates neurological hypotheses validated in molecular, organoid, and clinical systems  (2512.13724 - Noori et al., 13 Dec 2025) in Results, Section “Proton links Parkinson's genetics and molecular data”