Objective: We aim to evaluate the performance of a Parkinson’s disease (PD) polygenic risk score (PRS), estimated using genome-wide association (GWAS) data from a European cohort, in a Latino PD cohort. [1]

Background: PD risk prediction can now be improved by the inclusion of aggregated PD GWAS summary data in the form of a PRS. However, PD GWAS data have an overall European bias; as a consequence, the out-of-sample accuracy of PRS-based prediction might suffer when applied to non-Europeans, limiting its clinical utility.

Method: We calculated a PD-PRS for 1497 (807 cases) subjects from the Latin American Research Consortium on the Genetics of PD (LARGE-PD) using GWAS-significant summary statistics from Nalls et al. (2019). [2] We tested the PRS for association with PD using 10-fold cross-validation and logistic regression. Using Beagle 5.0, we inferred SNCA haplotypes by phasing a merged dataset of LARGE-PD, 1000 Genomes Project, and the Peruvian Genome Project genotype data.

Results: The European-derived PRS is significantly associated with PD (p-value < 2×10-16), with an area under the receiver-operator curve (AUC) of 0.669 (95% CI: 0.642-0.697). This exceeds the AUC of 0.651 found in a European cohort (Nalls et al. 2019). Peruvian subjects drive this result, with an AUC of 0.687 (95% CI: 0.644-0.730) compared to an AUC of only 0.617 (95% CI: 0.578-0.655) in the remaining recruitment sites. A major contributor to this discrepancy is rs356182 in SNCA. This variant has the largest beta coefficient out of the common PD risk variants and is enriched in Peruvian cases compared to expectations. We have identified two major rs356182 G-allele haplotypes; one haplotype represents 79.3% of European carriers, while the other haplotype represents 75.1% of East Asian carriers. In LARGE-PD, only the second haplotype is associated with PD risk (p-value = 0.00272).

Conclusion: The performance of the European-derived PD PRS exceeds expectations for a subset of our cohort but underperforms in the remainder, demonstrating a need to improve overall understanding of cross-population PRS-based risk prediction prior to its clinical implementation. We also found evidence for differential PD risk between the two major haplotypes carrying the rs356182 G allele, warranting further exploration.

References: [1] Submitted on behalf of the Latin American Research Consortium on the Genetics of PD (LARGE-PD) [2] Nalls MA, Blauwendraat C, Vallerga CL, et al. Identification of novel risk loci, causal insights, and heritable risk for Parkinson’s disease: a meta-analysis of genome-wide association studies. Lancet Neurol. 2019;18(12):1091-1102. doi:10.1016/S1474-4422(19)30320-5

« Back to MDS Virtual Congress 2020

MDS Abstracts - https://www.mdsabstracts.org/abstract/parkinsons-disease-polygenic-risk-prediction-and-haplotype-analysis-in-a-south-american-cohort/