Lysosomal dysfunction has been implicated in the pathogenesis of Parkinson’s disease (PD). Genetic variation in lysosomal pathway genes may contribute to PD susceptibility, motivating development of a lysosomal polygenic risk score (lyso-PRS) to capture cumulative genetic risk from lysosomal loci. Prior to this study, lyso-PRS had not been extensively evaluated across multiple ancestries or compared between idiopathic PD, GBA1-PD, non-manifesting GBA1 carriers (NMC), and healthy controls.
The study had three main objectives:
Investigators analyzed Neurobooster array genotype data from the Global Parkinson’s Genetics Program (GP2). The primary dataset comprised Tier 2 GP2 release 11 and included 32,054 idiopathic PD cases and 18,188 healthy controls across ten ancestries. A European test cohort from GP2 served for primary analyses, and an independent European cohort (European-TUEPAC) was used for validation. The lyso-PRS was derived from lysosomal pathway variants (details of score construction and variant selection are reported in the study’s clinical protocols and code repository). Statistical performance metrics included area under the receiver operating characteristic curve (AUC) and p values; associations with continuous clinical outcomes were reported as regression beta coefficients and p values.
Overall, the lyso-PRS demonstrated variable predictive performance by ancestry. The strongest performance was observed in the European test cohort (European-GP2: AUC = 0.602, p = 2.02e-172) and this was supported by validation in the independent European-TUEPAC cohort (AUC = 0.632, p = 5.34e-20). Performance in other ancestral groups was lower, with reported AUCs ranging from 0.51 to 0.59 and statistically significant p values (p < 0.05). These results indicate substantially better discriminative ability in people of European ancestry compared with non-European groups using the current score construction and data.
Within the European dataset, the lyso-PRS differentiated GBA1-PD cases from healthy controls (AUC = 0.621, p = 7.83e-71). The lyso-PRS also distinguished GBA1-PD from non-manifesting GBA1 carriers (NMC) with identical reported AUC (AUC = 0.621, p = 1.05e-18). These findings suggest that lysosomal genetic burden, as summarized by the lyso-PRS, is higher in manifest GBA1-PD relative to both unaffected carriers and population controls in Europeans.
In the European-GP2 cohort, higher lyso-PRS was associated with clinical measures of earlier disease and greater severity. Specifically, a higher lyso-PRS correlated with earlier age at onset (β = −0.21, p = 0.022). Higher lyso-PRS values were also associated with higher Unified Parkinson’s Disease Rating Scale (UPDRS) scores, with reported beta coefficients in the range 0.32–1.13 and all p values < 0.05. The authors note that not all clinical associations observed in the European-GP2 cohort were replicated in the separate validation cohort, indicating inconsistency for some findings.
The study concludes that a lysosomal polygenic risk score (lyso-PRS) is associated with both idiopathic PD and GBA1-PD status and shows its strongest predictive performance in individuals of European ancestry. Reduced performance in other ancestries highlights the need to improve score accuracy and validation in diverse populations. Associations between lyso-PRS and clinical features in Europeans—earlier age at onset and higher UPDRS scores—suggest the lyso-PRS may capture clinically relevant genetic burden, but incomplete replication emphasizes the need for further investigation.
Implications include the necessity for larger, ancestrally diverse datasets and prospective validation to refine lyso-PRS construction, improve cross-ancestry transferability, and determine clinical utility for risk stratification or prognosis.
Data used in the study were obtained from the Global Parkinson’s Genetics Program (GP2) Tier 2 release 11 (accessible via GP2/Zenodo and AMP PD). The European Tuebingen Parkinson cohort data are available in anonymized form on request. The authors report that written informed consent was obtained from all participants and that local ethics approvals were secured at collaborating sites; oversight and consent procedures were reviewed by GP2 governance. No competing interests were declared. The study’s clinical protocols and code are available in the cited repository.