Parkinson’s disease genetics across diverse ancestries: an observational genetic study of causal and risk variants with translational implications

Lara M Lange; Zih-Hua Fang; Mary B Makarious; Nicole Kuznetsov; Kajsa Atterling Brolin; Shannon Ballard; Prof Soraya Bardien; Maria Leila Doquenia; Prof Peter Heutink; Prof Henry Houlden; Hirotaka Iwaki; Simona Jasaityte; Lietsel Jones; Johanna Junker; Rauan Kaiyrzhanov; Mathew J Koretsky; Kishore R Kumar; the Latin American Research Consortium on the Genetics of Parkinson's Disease (LARGE-PD); Hampton L Leonard; Kristin S Levine; Prof Shen-Yang Lim; Niccoló E Mencacci; Prof Wael M Y Mohamed; Mike A Nalls; Prof Alastair J Noyce; Rajeev Ojha; Prof Njideka U Okubadejo; Shoaib ur Rehman; Laurel Screven; Chingiz Shashkin; Sophia Sopromadze; Eleanor J Stafford; Prof Ai Huey Tan; Manuela Tan; Zaruhi Tavadyan; Joanne Trinh; Bayasgalan Tserensodnom; Prof Enza Maria Valente; Dan Vitale; Nazira Zharkinbekova; Prof Katja Lohmann; Sara Bandres-Ciga; Cornelis Blauwendraat; Andrew B Singleton; Prof Huw R Morris; Prof Christine Klein; the Global Parkinson’s Genetics Program (GP2)

doi:10.1016/S1474-4422(26)00198-5

. Author manuscript; available in PMC: 2026 Aug 7.

Published in final edited form as: Lancet Neurol. 2026 Aug;25(8):741–754. doi: 10.1016/S1474-4422(26)00198-5

Parkinson’s disease genetics across diverse ancestries: an observational genetic study of causal and risk variants with translational implications

Lara M Lange ^1,², Zih-Hua Fang ³, Mary B Makarious ^4,⁵, Nicole Kuznetsov ^4,⁵, Kajsa Atterling Brolin ^6,⁷, Shannon Ballard ^4,⁵, Soraya Bardien ⁸, Maria Leila Doquenia ^9,^10,¹¹, Peter Heutink ¹², Henry Houlden ¹³, Hirotaka Iwaki ^4,⁵, Simona Jasaityte ¹⁴, Lietsel Jones ^4,⁵, Johanna Junker ^15,¹⁶, Rauan Kaiyrzhanov ^13,¹⁷, Mathew J Koretsky ^4,⁵, Kishore R Kumar ^18,^19,²⁰; the Latin American Research Consortium on the Genetics of Parkinson's Disease (LARGE-PD), Hampton L Leonard ^4,⁵, Kristin S Levine ^4,⁵, Shen-Yang Lim ²¹, Niccoló E Mencacci ²², Wael M Y Mohamed ^23,²⁴, Mike A Nalls ^4,⁵, Alastair J Noyce ⁷, Rajeev Ojha ²⁵, Njideka U Okubadejo ²⁶, Shoaib ur Rehman ²⁷, Laurel Screven ¹², Chingiz Shashkin ^28,²⁹, Sophia Sopromadze ^30,³¹, Eleanor J Stafford ¹⁴, Ai Huey Tan ²¹, Manuela Tan ^7,³², Zaruhi Tavadyan ³³, Joanne Trinh ², Bayasgalan Tserensodnom ³⁴, Enza Maria Valente ^35,³⁶, Dan Vitale ^4,⁵, Nazira Zharkinbekova ¹⁷, Katja Lohmann ², Sara Bandres-Ciga ⁵, Cornelis Blauwendraat ^12,³⁷, Andrew B Singleton ¹², Huw R Morris ^14,^38,^#, Christine Klein ^2,^#; the Global Parkinson’s Genetics Program (GP2)

¹Laboratory of Neurogenetics, National Institute on Aging, Bethesda, MD, USA

²Institute of Neurogenetics, University of Luebeck, Luebeck, Germany

³German Center for Neurodegenerative Diseases (DZNE), Tuebingen, Germany

⁴DataTecnica, Washington, DC, USA

⁵Center for Alzheimer's and Related Dementias (CARD), National Institute on Aging and National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD, USA

⁶Translational Neurogenetics Unit, Department of Experimental Medical Science, Lund University, Lund, Sweden

⁷Centre for Preventive Neurology, Wolfson Institute of Population Health, Queen Mary University of London, London, UK

⁸Division of Molecular Biology and Human Genetics, Faculty of Medicine and Health Sciences, Stellenbosch University; South African Medical Research Council/Stellenbosch University Genomics of Brain Disorders Research Unit, Cape Town, South Africa

⁹Department of Neuroscience and Brain Health, Metropolitan Medical Center, Manila, Philippines

¹⁰Department of Clinical Neurosciences, Mary Mediatrix Medical Center, Lipa, Batangas, Philippines

¹¹Department of Anatomy, University of the East Ramon Magsaysay Memorial Medical Center, Quezon City, Philippines

¹²The Global Parkinson’s Genetics Program (GP2), Chevy Chase, MD, USA

¹³University College London, Institute of Neurology, Department of Neuromuscular Diseases, Queen Square, London, UK

¹⁴Department of Clinical and Movement Neurosciences, UCL Queen Square Institute of Neurology, London, UK

¹⁵University Hospital Schleswig Holstein, Campus Luebeck, Luebeck, Germany

¹⁶Center for Movement Disorders and Neuromodulation, University Hospital Duesseldorf, Duesseldorf, Germany

¹⁷Department of Neurology, South Kazakhstan Medical Academy, Shymkent, Kazakhstan

¹⁸Neurology Department and Molecular Medicine Laboratory, Concord Repatriation General Hospital and University of Sydney, Concord, NSW, Australia

¹⁹Translational Neurogenomics Group, Genomics and Inherited Disease Program, The Garvan Institute of Medical Research, Darlinghurst, NSW, Australia

²⁰Faculty of Medicine and Health, University of New South Wales, Sydney, NSW, Australia

²¹Division of Neurology, Department of Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia

²²Department of Neurology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA

²³Department of Basic Medical Sciences, Kulliyyah of Medicine, International Islamic University Malaysia (IIUM), Kuantan, Pahang, Malaysia

²⁴Clinical Pharmacology Department, Menoufia Medical School, Menoufia University, Shebin El-Kom, Menoufia, Egypt

²⁵Department of Neurology, Tribhuvan University Teaching Hospital, Kathmandu, Nepal

²⁶Neurology Unit, Department of Medicine, College of Medicine, University of Lagos, Lagos State, Nigeria

²⁷Department of Biotechnology, University of science and Technology Bannu, Pakistan

²⁸International Research Institute of Postgraduate Education, Department of Neurosurgery and Neurology, Almaty, Kazakhstan

²⁹Shashkin Clinic, Almaty, Kazakhstan

³⁰Department of Neurology, Ivane Javakhishvili Tbilisi State University, Tbilisi, Georgia

³¹East European University (EEU), Tbilisi, Georgia

³²Department of Neurology, Oslo University Hospital, Oslo, Norway

³³National Institute of Health, Yerevan, Armenia

³⁴Department of Neurology, School of Medicine, Mongolian National University of Medical Sciences, Ulaanbaatar, Mongolia

³⁵Department of Molecular Medicine, University of Pavia, Pavia, Italy

³⁶Neurogenetics Research Center, IRCCS Mondino Foundation, Pavia, Italy

³⁷Coalition for Aligning Science, Chevy Chase, MD, USA

³⁸UCL Movement Disorders Centre, University College London, London, UK

^✉

Correspondence to Lara M. Lange, Laboratory of Neurogenetics, National Institute on Aging (NIA), National Institutes of Health, (NIH); 9000 Rockville Pike, Bethesda, MD 20892, United States of America and Institute of Neurogenetics, University of Luebeck; Ratzeburger Allee 160, 23538, Luebeck, Germany, lara.lange@nih.gov; or Christine Klein, Institute of Neurogenetics, University of Luebeck; Ratzeburger Allee 160, 23538 Luebeck, Germany, christine.klein@uni-luebeck.de

Shared last authors

Author contributions

LML, ZHF, and NK accessed and verified the data reported in this manuscript. LML, SoB, HH, RK, KRK, SYL, NEM, WMYM, AJN, RO, NUO, SUR, CS, SS, AHT, ZT, BT, EMV, NZ, HRM, and CK carried out study assessments for participants, including clinical data and biospecimen collection. ZHF, MBM, MJK, HLL, KSL, MAN, and DV were responsible for processing, performing quality control, and harmonizing genotyping and sequencing data from the Global Parkinson’s Genetics Program. HI and LJ were responsible for processing and harmonizing clinical data from the Global Parkinson’s Genetics Program. KAB, ShB, MLD, SJ, JJ, EJS, and MT contributed to the acquisition of data. MBM, HLL, KL, SBC, CB, and AS contributed to the analysis of genetic data. LML and CK conceptualised the study. PH, LS, and JT supported the execution of the study. CB, ABS, HRM, and CK supervised the study. LML drafted the manuscript. All authors had full access to all the data in the study and had final responsibility for the decision to submit the manuscript for publication. LML, CB, ABS, HRM, and CK were responsible for the decision to submit the manuscript.

PMCID: PMC13446284 NIHMSID: NIHMS2198437 PMID: 42456684

Abstract

Background:

The genetic architecture of Parkinson's disease, encompassing the distribution, frequency, and effect of disease-associated genetic variants, varies considerably across ancestries, yet most genetic studies in Parkinson’s disease have focused on European-ancestry individuals. We aimed to characterize the distribution of established Parkinson’s disease causal variants as well as risk-associated variants with clinical implications, i.e., variants in genes involved in pathways targeted by ongoing clinical trials, across ancestrally diverse populations.

Methods:

We conducted a multi-ancestry observational, cross-sectional genetic study using retrospective data from GP2 Release 11 (released in December 2025). The study investigated causal and risk variants, including copy number variants, in established Parkinson’s disease and parkinsonism-associated genes, following the recommendations of the MDS Task Force on the Nomenclature of Genetic Movement Disorders, including GBA1, LRRK2, SNCA, VPS35, RAB32, PINK1, PRKN, PARK7, ATP13A2, DCTN1, DNAJC6, FBXO7, JAM2, RAB39B, SLC20A2, SYNJ1, VPS13C, and WDR45. PD cases were diagnosed based on established clinical criteria, including the Parkinson’s UK Brain Bank and/or Movement Disorder Society diagnostic criteria, and control participants were defined as individuals without evidence of neurodegenerative disease and unrelated to PD participants included in the study. We analysed genome and exome sequencing and genotyping data of 99,783 individuals, including 58,559 individuals with PD and 41,224 controls, from eleven different genetically inferred ancestries (i.e., African, African-Admixed, Ashkenazi Jewish, Latino, Central Asian, Complex Admixture, East Asian, European, Finnish, Middle Eastern, and South Asian), defined using reference population-based ancestry inference methods. We calculated allele frequencies for all investigated variants in PD cases and controls, both overall and stratified by ancestry.

Findings:

Approximately 29% (29001/99783) of individuals (~26% [15443/58559] of individuals with Parkinson’s disease, and ~33% [13558/41224] of controls) were from underrepresented populations, i.e., non-European and non-Ashkenazi Jewish. Our study revealed both shared genetic contributors across ancestries as well as ancestry-specific differences in variant frequencies and the spectrum of variants within PD-associated genes. Overall, ~2% (1217/58559) of individuals with Parkinson’s disease carried a causal variant, with substantial variations across ancestries ranging from ~0·5% (10/2844 and 3/550) in African and African-Admixed to ~7% (92/1347) in Middle Eastern and >10% (251/2343) Ashkenazi Jewish ancestries. Risk variants in GBA1 and LRRK2 were identified in 11·8% (6893/58559) of individuals with Parkinson’s disease and 8·7% (3578/41224) of controls. GBA1 risk variants were most frequent overall and identified across all ancestries, but variant frequency and spectra differed substantially between ancestries, i.e., frequencies ranged from ~4% (195/4773) in East Asians to ~53% (1505/2844) in African ancestry. Similarly, LRRK2 causal and risk variants showed ancestry-specific enrichment, with highest frequencies of causal variants in Ashkenazi Jewish (250/2343, 10.7%) and Middle Eastern (59/1347, 4.4%) populations, while risk variants were predominantly identified in East Asian ancestry (601/4773, 12.6%). Carriers of causal variants in PRKN, commonly including deletions and duplications, were also identified across all ancestries except Ashkenazi Jewish, with the highest frequency of 1.3% (17/1347) in Middle Eastern ancestry, while frequencies in all other ancestries were below 1%.

Interpretation:

This large-scale, multi-ancestry genetic study offers crucial insights into the population-specific genetic architecture of PD. Whereas clinical trials targeting GBA1 and LRRK2 variant carriers are currently primarily performed in Europe and the US, this study underscores the critical need for increased ancestral diversity in PD research to improve diagnostic accuracy, enhance our understanding of disease mechanisms across populations, and ensure equitable application of and access to emerging genetically-informed therapies.

Funding:

Aligning Science Across Parkinson’s (ASAP) through the Global Parkinson's Genetics Program (GP2)

Keywords: Parkinson’s Disease, Multi-ancestry, Genetics, LRRK2, GBA1, SNCA, PRKN, causal variants, risk variants, precision medicine

Introduction

Genetic discoveries have transformed our understanding of Parkinson’s disease (PD). Genetic factors substantially contribute to PD risk and progression.¹ The genetic factors range from rare variants with large effect sizes to common variants with moderate effect sizes. More than 130 independent risk loci have been identified by genome-wide association studies (GWAS),² and over 15 genes have been linked to monogenic PD and parkinsonism.^1,3

A critical gap in PD genetics research is the lack of ancestral diversity.⁴ Although genetic forms of PD are found in all regions across the world, approximately 75% of all genetic PD studies to date focused on European-ancestry individuals.^5,6 Research in diverse populations is critical to uncovering unique genetic contributions to disease and improve diagnosis, risk assessment, and genetic counselling. This need for greater ancestral diversity in PD genetics research is particularly relevant for clinical trials driven by genetic discoveries, e.g., targeting glucocerebrosidase or LRRK2 kinase (encoded by the PD-linked genes GBA1 and LRRK2, respectively).⁷ The absence of diversity in genetic studies may limit the broader applicability of emerging opportunities, as it is unclear whether current genetic targets are relevant across ancestries or whether population-specific variants exist.

The Global Parkinson’s Genetics Program (GP2; https://gp2.org/)⁸ is a large-scale initiative actively addressing this gap by investigating the genetic underpinnings of PD and parkinsonism across ancestry-diverse populations.^5,9 As part of these efforts, several GWAS were performed in populations historically underrepresented in genetic studies, including African¹⁰, Latin American¹¹, and South Asian¹², leading to the identification of both shared and ancestry-specific risk loci. In parallel, studies in diverse populations have identified marked ancestry enrichment of specific disease-related variants, such as LRRK2 p.R1067Q in East Asia ¹³ and GBA1 p.K198E in Colombia ¹⁴.

These observations highlight the need for systematic, large-scale, multi-ancestry investigations. Accordingly, the aim of our study was to comprehensively and descriptively characterize the carrier frequencies, and ancestry-specific presence and spectrum of causal and risk variants across diverse populations, and to evaluate their translational relevance for genetic screening and genetically informed therapy.

Methods

Study design and participants

Our study workflow is displayed in Figure 1. We analysed short-read genome sequencing, clinical exome sequencing and Illumina NeuroBooster Array genotyping data from GP2’s Release 11 (https://doi.org/10.5281/zenodo.17753486), comprising all data available and processed within GP2 up to December 2025. A detailed data overview is provided in Supplementary Table 1. The total sample comprised 99,783 individuals, including 58,559 individuals with PD and 41,224 healthy controls. This analysis includes participant data from 144 different cohorts at 129 sites. Our analyses included a subset of previously published individuals from PD GENEration (n=8902)¹⁵ and ROPAD (n=3113)¹⁶, now both part of the harmonised GP2 dataset.

All participants in GP2 underwent neurological assessment according to local study protocols. A diagnosis of PD was based on established clinical criteria, including the Parkinson’s UK Brain Bank and/or Movement Disorder Society diagnostic criteria. Control participants were defined as individuals without evidence of neurodegenerative disease and unrelated to participants with Parkinson’s disease included in the study.

This study was conducted in accordance with the ethical standards of the institutional and national research committees and approved by the ethics committees or institutional review boards at all GP2 sites that provided samples and data for this study. Informed consent for study participation was obtained from all participants by each local site. All cohorts recruited to the GP2 initiative underwent a thorough review of the consent forms in the Operations and Compliance working group, ensuring that each contributing study abides by the ethics guidelines set out by their institutional review boards.

Procedures

Data processing and quality control were performed using GenoTools, as described elsewhere¹⁷; details are provided in the Supplementary Material. Genetic ancestry was inferred using GenoTools, which leverages high-quality variants shared with curated reference panels from the 1000 Genomes Project, Human Genome Diversity Project, and Ashkenazi reference panels. Reference variants have undergone quality control for minor allele frequency, missingness, Hardy-Weinberg equilibrium, and linkage disequilibrium. Based on this pipeline, samples were divided into the following ancestries: African-Admixed (AAC; 550 PD, 881 controls), African (AFR; 2844 PD, 4489 controls), Ashkenazi Jewish (AJ; 2343 PD, 913 controls), Latinos and Indigenous People of the Americas (AMR; 2809 PD, 1654 controls), Complex Admixture (CAH; 902 PD, 444 controls), Central Asian (CAS; 1284 PD, 1480 controls), East Asian (EAS; 4773 PD, 2809 controls), European (EUR; 40773 PD, 26753 controls), Finnish (FIN; 127 PD, 22 controls), Middle Eastern (MDE; 1347 PD, 1339 controls), and South Asian (SAS; 807 PD, 440 controls). Because publicly available reference samples are limited for certain populations and highly admixed groups, GenoTools incorporates an approach to identify such individuals and assign them to a distinct category not represented in the reference panel, termed “Complex Admixture History” (CAH). Since ancestry-specific interpretation is limited for these individuals, this group is not discussed in the Results, but all corresponding cohort information and results for this ancestry group are provided in the Tables and Supplementary Materials.

We focused on variants in genes linked to PD and parkinsonism following the recommendations of the MDS Task Force on the Nomenclature of Genetic Movement Disorders³, including variants in i) LRRK2, SNCA, and VPS35 linked to autosomal-dominant PD, ii) PARK7, PINK1, and PRKN linked to autosomal-recessive PD, and iii) ATP13A2, DCTN1, DNAJC6, FBXO7, JAM2, RAB39B, SLC20A2, SYNJ1, VPS13C, and WDR45 linked to atypical or complex parkinsonism. We included only PARK-designated genes, i.e., all genes confidently linked to PD and parkinsonism based on the available evidence from the literature and following criteria defined by the MDS Task Force; genes linked to combined phenotypes including PD or parkinsonism (e.g., dystonia-parkinsonism, Frontotemporal Dementiaparkinsonism, etc.) were beyond the scope of this study. We added RAB32 p.S71R, identified as causal for PD after the Task Force recommendations were published. We included pathogenic/likely pathogenic (according to ClinVar; https://www.ncbi.nlm.nih.gov/clinvar/) variants in these genes and also investigated PD risk-associated variants in GBA1 and LRRK2. For details on pathogenicity evaluation, see Supplementary Material (appendix pp. 4-5 and pp. 22-27). A full list of all identified and investigated variants, including their pathogenicity evaluation and detectability by genotyping or sequencing, is provided in Supplementary Table 2. Copy number variation (CNV) analyses for SNCA and PRKN were performed for samples with available genotyping data as described previously.¹⁸ Carriers of two pathogenic variants in recessive genes and an age at onset ≤60 years were considered likely compound heterozygous without further validation, based on findings from previous studies.¹⁹

Statistical analysis

We calculated allele frequencies for all investigated variants in PD cases and controls, both overall and stratified by ancestry. For selected variants meeting predefined criteria, we performed single-variant association analyses using logistic regression. Variants were included if they met the following criteria: (1) available from imputed genotyping data, (2) observed in ≥2 ancestry groups, and (3) ≥1 carrier per group (cases and controls) across these ancestry groups. We used sex and the first five principal components as covariates. Age was not included as a covariate because it was missing for a considerable proportion of individuals; sensitivity analyses including age in the subset of samples with available data yielded comparable results to analyses performed without age as a covariate. Analyses were conducted using PLINK v2.

Role of the funding source

The funders of this research had no role in data collection, analysis, interpretation, writing of the manuscript and the decision to submit.

Results

Sample characteristics are summarized in Table 1. Across all ancestries, 2·1% (1217/58559) of individuals with PD carried a causal and 11·8% (6893/58559) a risk-associated variant (Table 2). In comparison, causative variants were observed in 0.3% (108/41224) and risk variants in 8·7% (3578/41224) of controls (Table 2).

Table 1.

Cohort characteristics.

	PD					Healthy controls
Ancestry	Samples (n)	Male (%)	Median Age^* (IQR), range	Median AAO^** (IQR), range	Positive FH of PD^*** (%)	Samples (n)	Male (%)	Median Age^* (IQR), range	Positive FH of PD^*** (%)
AAC	550	334 (60.7)	67 (60-73), 29-95 [unk for n = 86]	60 (51-68), 13-90 [unk for n = 147]	75 (13.6), [unk for n = 102]	881	314 (35.6)	65 (58-71), 18-91 [unk for n = 39]	3 (0.3), [unk for n = 450]
AFR	2844	1983 (69.7)	65 (57-72), 22-96 [unk for n = 177]	58 (54-66), 12-92 [unk for n = 182]	326 (11.5), [unk for n = 1006]	4489	2303 (51.3)	63 (56-69), 18-99 [unk for n = 71]	11 (0.2), [unk for n = 1878]
AJ	2343	1570 (67.0)	71 (64-77), 32-94 [unk for n = 182]	64 (56-71), 12-93 [unk for n = 466]	578 (24.7), [unk for n = 385]	913	530 (58.1)	67 (60-74), 23-95 [unk for n = 284]	31 (3.4), [unk for n = 586]
AMR	2809	1618 (57.6)	64 (54-71), 18-97 [unk for n = 192]	54 (44-63), 10-92 [unk for n = 1689]	485 (17.3), [unk for n = 793]	1654	582 (35.2)	59 (53-65), 18-92 [unk for n = 71]	84 (5.1), [unk for n = 210]
CAH	902	527 (58.4)	63 (54-70), 18-92 [unk for n = 151]	55 (45-64), 13-88 [unk for n = 313]	164 (18.2), [unk for n = 245]	444	184 (41.4)	55 (32-66), 18-94 [unk for n = 53]	5 (1.1), [unk for n = 321]
CAS	1284	599 (46.7)	62 (54-69), 18-94 [unk for n = 464]	53 (43-61), 14-89 [unk for n = 667]	108 (84.), [unk for n = 635]	1480	612 (41.4)	59 (52-65), 29-96 [unk for n = 689]	2 (0.1), [unk for n = 1454]
EAS	4773	2530 (53.0)	67 (58-73), 19-96 [unk for n = 671]	56 (47-65), 10-92 [unk for n = 1078]	943 (19.8), [unk for n = 625]	2809	1890 (67.3)	63 (55-70), 20-98 [unk for n = 1166]	11 (0.4), [unk for n = 1737]
EUR	40773	25633 (62.9)	68 (60-75), 19-100 [unk for n = 6052]	60 (51-68), 10-98 [unk for n = 8635]	8300 (20.4), [unk for n = 15436]	26753	13148 (49.1)	64 (57-71), 18-99 [unk for n = 4507]	1013 (3.8), [unk for n = 19914]
FIN	127	59 (46.5)	67 (58-73), 38-86 [unk for n = 59]	57 (50-66), 35-76 [unk for n = 60]	20 (15.7), [unk for n = 77]	22	11 (50.0)	72 (64-78), 49-85 [unk for n = 6]	1 (4.5), [unk for n = 16]
MDE	1347	797 (59.2)	64 (56-70), 21-94 [unk for n = 339]	53 (44-62), 11-85 [unk for n = 485]	337 (25.0), [unk for n = 485]	1339	603 (45.0)	63 (53-70), 22-96 [unk for n = 601]	0 (0), [unk for n = 1263]
SAS	807	528 (65.4)	63 (53-71), 21-91 [unk for n = 68]	56 (45-64), 16-85 [unk for n = 93]	207 (25.7), [unk for n = 131]	440	304 (69.1)	57 (50-63), 19-84 [unk for n = 92]	1 (0.2), [unk for n = 228]
Total	58559	36178 (61.8)	67 (59-74), 18-100 [unk for n = 8437]	60 (50-67), 10-98 [unk for n = 13791]	11543 (19.7), [unk for n = 19893]	41224	20481 (49.7)	63 (56-70), 18-99 [unk for n = 7579]	1162 (2.8), [unk for n = 28057]

Open in a new tab

Age refers to the age at recruitment/age at sample collection. For controls, individuals with an age <18 years or >100 years were considered outliers and excluded from these statistics. For PD, individuals with an age <10 years or >100 years were considered outliers and excluded from these statistics.

^**

Individuals with an AAO <10 years or >100 years were considered outliers and excluded from these statistics.

^***

This includes related individuals.

AAO = Age at onset, FH = Family history, IQR = Interquartile range, NA = not available/not applicable, PD = Parkinson's disease, unk = unknown

Table 2.

Summary of genetic findings and yields across ancestries.

Genetic ancestry	Group	Investigated samples (n)	PD causal variants									PD risk variants^*			Dual carriers ^**	TOTAL risk and causal variant carriers, yield (%)
			Typical autosomal-dominant PD				Early-onset recessive PD			Atypical PD genes	TOTAL causal variant carriers, yield (%)	Risk-associated		TOTAL risk variant carriers, yield (%)
			LRRK2	SNCA	VPS35	RAB32	PINK1	PRKN	PARK7	Atypical PD genes	TOTAL causal variant carriers, yield (%)	GBA1	LRRK2	TOTAL risk variant carriers, yield (%)
AAC	PD	550	1	0	0	0	0	1	0	1	3 (0.5)	171 / 3 / 24	0	198 (36.0)	2	199/550 (36.2)
AAC	Controls	881	3	0	0	0	0	0	0	0	3 (0.3)	201 / 3 / 6	1	211 (24.0)	1	213/881 (24.2)
AFR	PD	2844	3	2	0	0	1	4	0	0	10/2844 (0.4)	1133 / 33 / 339	2	1507/2844 (53.0)	7	1510/2844 (53.1)
AFR	Controls	4495	3	0	0	0	0	0	0	1	4/4489 (0.1)	1503 / 3 / 183	1	1690/4489 (37.6)	3	1691/4489 (37.7)
AJ	PD	2343	245 / 5	1	0	0	0	0	0	0	251/2343 (10.7)	383 / 6 / 5	0	394/2343 (16.8)	25	620/2343 (26.5)
AJ	Controls	913	20	0	0	0	0	0	0	0	20/913 (2.2)	73 / 1 / 2	0	76/913 (8.3)	1	95/913 (10.4)
AMR	PD	2809	59	2	0	0	0	24	0	0	85/2809 (3.0)	151 / 8 / 3	0	162/2809 (5.8)	2	245/2809 (8.7)
AMR	Controls	1654	0	0	0	0	0	0	0	0	0/1654 (0)	30 / 0 / 0	0	30/1654 (1.8)	0	30/1654 (1.8)
CAH	PD	902	21	1	1	0	0	12	0	0	35/902 (3.9)	130 / 2 / 7	9	148/902 (16.4)	4	179/902 (19.8)
CAH	Controls	444	2	0	0	0	0	0	0	0	2/444 (0.5)	40 / 1 / 2	3	46/444 (10.4)	0	48/444 (10.8)
CAS	PD	1284	1	0	0	1	0	7	0	0	9/1284 (0.7)	73 / 1 / 1	23	98/1284 (7.6)	2	105/1284 (8.2)
CAS	Controls	1480	0	0	0	0	0	1	0	0	1/1480 (0.1)	29 / 0 / 0	14	43/1480 (2.9)	0	44/1480 (3.0)
EAS	PD	4773	15	4	5	0	17	17	0	1	69/4773 (1.44)	190 / 5 / 0	574 / 16 / 11	796/4773 (16.7)	25	840/4773 (17.6)
EAS	Controls	2809	2	0	0	0	0	0	0	0	2/2809 (0.1)	14 / 1 / 0	165 / 4 / 1	185/2809 (6.6)	0	187/2809 (6.7)
EUR	PD	40773	451 / 2	31	8	17	5	126	3	8	651/40773 (1.6)	3312 / 99 / 20	21	3452/40773 (8.5)	49	4054/40773 (9.9)
EUR	Controls	26761	65	2	0	0	0	2	0	2	71/26753 (0.3)	1249 / 11 / 2	6	1268/26753 (4.7)	6	1333/26753 (5.0)
FIN	PD	127	0	0	0	0	0	1	0	0	1/127 (0.8)	25 / 1 / 0	0	26/127 (20.5)	1	26/127 (20.5)
FIN	Controls	22	0	0	0	0	0	0	0	0	0/22 (0)	4 / 0 / 0	0	4/22 (18.2)	0	4/22 (18.2)
MDE	PD	1347	56 / 3	1	0	4	11	17	0	0	92/1347 (6.8)	70 / 1 / 2	0	73/1347 (5.4)	7	158/1347 (11.7)
MDE	Controls	1340	4	0	0	0	0	1	0	0	5/1339 (0.4)	20 / 0 / 0	0	20/1339 (1.5)	0	25/1339 (1.9)
SAS	PD	807	1	0	1	0	4	5	0	0	11/807 (1.4)	38 / 0 / 1	0	39/807 (4.8)	1	49/807 (6.1)
SAS	Controls	440	0	0	0	0	0	0	0	0	0/440 (0)	4 / 0 / 0	0	4/440 (0.9)	0	4/440 (0.9)
TOTAL	PD	58559	853 / 10	42	15	22	48	214	3	10	1217/58559 (2.1)	5676 / 159 / 402	629 / 16 / 11	6893/58559 (11.8)	125	7985/58559 (13.6)
TOTAL	Controls	41224	99	2	0	0	0	4	0	3	108/41224 (0.3)	3167 / 20 / 195	191 / 4 / 1	3578/41224 (8.7)	11	3675/41224 (8.9)

Open in a new tab

This table reports gene-specific numbers of variant carriers across ancestries (genetic findings), with the total reflecting the number of unique individuals. For autosomal dominant genes, counts reflect heterozygous carriers, whereas for autosomal recessive genes, counts reflect (likely) biallelic carriers. In columns with three different values, the first denotes heterozygous carriers, the second individuals carrying two different heterozygous variants, and the third homozygous variant carriers. For columns with two different values, the first denotes heterozygous and the second homozygous carriers. Further, this table summarizes the total counts and proportions (genetic yield) of individuals carrying risk variants only, causal variants only, or both. The combined category represents unique individuals with any qualifying variant. Because some individuals carry both risk and causal variants, combined counts and percentages are lower than the sum of category-specific values.

All pathogenic/likely pathogenic GBA1 variants are considered risk variants in the context of PD, including Gaucher disease-causing variants of all severities (e.g., severe and mild) as well as variants only associated with an increased risk of PD.

^**

Dual carriers refer to individuals harboring pathogenic/likely pathogenic or Parkinson’s disease risk variants in two different genes. The total number reflects the number of individuals rather than variants; therefore, dual carriers are counted only once in the total, and column-wise sums may exceed the total.

AAC = African admixed, AFR = African, AJ = Ashkenazi Jewish, AMR = Latinos and Indigenous people of the Americas, CAH = Complex admixture, CAS = Central Asian, EAS = East Asian, EUR = European, FIN = Finnish, MDE = Middle East, SAS = South Asian

We identified 113 distinct pathogenic single nucleotide variants (SNVs) in the 16 genes studied (including single heterozygous variants in recessive genes), alongside two LRRK2 risk variants, 95 GBA1 risk variants, and structural variations like SNCA multiplications and PRKN exon deletions and duplications (Supplementary Table 2; appendix pp. 22-27). The allele frequencies of all identified variants as well as number of variant carriers, stratified by phenotype and ancestries, are provided in Supplementary Table 3. Only PRKN deletions and duplications and three GBA1 SNVs (i.e., p.E365K, p.T408M, and p.L483P) were detected across all ancestries, whereas the remaining variants were only identified in some ancestries. All variants identified in healthy controls, along with their age ranges, are provided in Supplementary Table 4 (appendix pp. 29-31).

As expected, association analyses confirmed several well-established ancestry-specific associations (Supplementary Table 5; appendix pp. 32-34): GBA1 p.E365K and p.N409S were significantly associated with PD in AJ and EUR populations, while GBA1 p.T408M reached significance only in EUR. GBA1 rs3115534-G was significantly associated with PD in AFR and AAC populations. For LRRK2, significant associations were observed for p.R1628P and p.G2385R in EAS and for p.G2019S in EUR, AJ, and MDE populations.

The relative contribution of variants in different PD-associated genes in individuals with PD across ancestries is summarised in Table 2, Figure 2, and Supplementary Figure 1, illustrating both shared genetic contributors and ancestry-specific distributions of carriers across PD-associated genes. GBA1 variant carriers were most frequent overall and detected across all ancestries, albeit with substantially different variant spectra (Figure 3). Similarly, LRRK2 variants were identified across multiple ancestries with ancestry-specific variant profiles (Figure 4). Rarer PRKN SNVs and structural variations were also observed across most ancestries, whereas causative variants in other genes were only identified in certain populations. Variants in genes linked to atypical parkinsonism were only identified in a small subset of individuals (Supplementary Table 6; appendix p. 35).

Figure 2. — (A) Donut charts displaying the percentage of identified carriers per gene amongst all carriers per ancestry. Dual refers to carriers of variants in two different genes (e.g., *GBA1* and a causal gene). (B) Global map of clinical trial locations recruiting gene variant carriers (*GBA1, LRRK2, PINK1,* and *PRKN*) and targeting proteins encoded by PD-linked genes or pathways in which these genes are involved, obtained from the clinical trial registry https://clinicaltrials.gov/. Only one pin per gene and country is shown, regardless of the number of study sites within that country.

AAC = African Admixed, AFR = African, AJ = Ashkenazi Jewish, AMR = Latinos and Indigenous people of the Americas, CAS = Central Asian, EAS = East Asian, EUR = European, FIN = Finnish, MDE = Middle Eastern, SAS = South Asian.

Figure 3. — Donut charts illustrate the distribution of *GBA1* variants in individuals with Parkinson’s disease stratified by variant severity (inner ring: Severe, Mild, Risk, Unknown) and specific variants (outer ring) across ancestries. Variant severity was assigned using the GBA1-PD browser (https://pdgenetics.shinyapps.io/GBA1Browser/). The size of each segment reflects the number of variant carriers within each ancestry. Inner-ring colours denote severity categories: Severe (purple), Mild (blue), Risk (green), and Unknown (grey). Outer-ring segments represent individual variants, with colour shades corresponding to their severity classification. To improve readability, variants identified in only a single individual were grouped as “Other” for CAS, MDE, and SAS; for AAC, AFR, AMR, EAS, and EUR, variants identified in ≤5 individuals were grouped. Numbers within the circles indicate the total number of *GBA1* variants identified per ancestry; individuals carrying two different *GBA1* variants contributed to both variant counts.

AAC = African Admixed, AFR = African, AJ = Ashkenazi Jewish, AMR = Latinos and Indigenous people of the Americas, CAH = Complex Admixture, CAS = Central Asian, EAS = East Asian, EUR = European, FIN = Finnish, MDE = Middle Eastern, SAS = South Asian.

Figure 4. — **(A)** Schematic representation of the LRRK2 protein with functional domains. Boxes indicate locations of the identified *LRRK2* variants within their respective domains. Variants highlighted in bold have been previously shown to increase LRRK2 kinase activity. **(B)** Donut charts illustrating the ancestry-specific spectrum of *LRRK2* variants in individuals with Parkinson’s disease. Each segment represents a specific variant; the segment size is proportional to the number of individuals carrying that variant. If two variants were present in one individual, both variants were counted. Segment colours correspond to the LRRK2 protein domain in which the variant is located, as indicated in Panel A. Only ancestries with more than five variant carriers are shown.

AJ = Ashkenazi Jewish, AMR = Latinos and Indigenous people of the Americas, CAS = Central Asian, EAS = East Asian, EUR = European, MDE = Middle Eastern.

Individuals with GBA1 risk variant associated PD showed a significantly earlier median age at onset of 3 to 8 years across multiple ancestries compared to those with idiopathic PD (IPD) (Supplementary Figure 2A and Supplementary Table 7). The AAO between LRRK2-PD and IPD showed similar trends in some ancestries, with LRRK2-PD showing a 1- to 3-year earlier onset (Supplementary Figure 2A and Supplementary Table 7). Further, the AAO distribution among GBA1 and LRRK2 carriers showed heterogeneity across ancestries (Supplementary Figures 2B and C).

Among individuals with PD of African-Admixed ancestry (AAC), pathogenic or risk variants were identified in 36·2% (199/550). Almost all carried GBA1 variants (198/199, 99·5%), predominantly driven by the common intronic risk variant, rs3115534-G (Figure 3). One individual (0·2%) with early-onset PD (AAO 27 years) carried a homozygous PRKN deletion, and another one (0·2%) harboured both a LRRK2 causal variant and a GBA1 risk variant. In controls, GBA1 risk variants were observed in 23·8% (210/881), again largely attributable to the rs3115534-G risk variant; pathogenic or risk LRRK2 variants were identified in 0·5% (4/881).

Pathogenic or risk variants were detected in 53·1% (1510/2844) of individuals with PD of African ancestry (AFR). GBA1 variants accounted for the majority of genetic findings, again predominantly driven by rs3115534-G (Figure 3). Less frequent findings included variants in LRRK2 (5/2844, 0·2%) and SNCA (2/2844, <0·1%), as well as variants in PRKN (4/2844, 0·1%) and PINK1 (1/2844, <0·1%) observed in individuals with early AAO (ranging from 27-54 years); most of these individuals (7/12) carried a co-occurring GBA1 variant. In controls, genetic findings were identified in 37·7% (1691/4489), predominantly in GBA1, largely attributable to rs3115534-G, with pathogenic or risk LRRK2 variants observed in a few individuals (4/4489, <0·1%).

In Ashkenazi Jewish ancestry (AJ), pathogenic or risk variants were identified in 26·5% (620/2343) of individuals with Parkinson’s disease. Almost all genetic findings involved GBA1 and LRRK2, each characterised by a predominant recurrent variant (GBA1 p.N409S and LRRK2 p.G2019S), alongside a small number of additional variants (Figures 3 and 4). Variants in GBA1 were observed in 394 individuals (16·8%) and LRRK2 variants in 250 individuals (10·7%); 25 of these individuals were dual GBA1-LRRK2 variant carriers. Only one PD individual carried a SNCA multiplication (<0·1%). Among controls, 10·4% (95/913) carried pathogenic and risk variants in GBA1 and LRRK2.

Among Latinos and Indigenous people of the Americas (AMR), 8·7% (245/2809) of individuals with Parkinson’s disease carried pathogenic and risk variants across several PD-linked genes, including GBA1 (162/2809, 5·8%), LRRK2 (59/2809, 2·1%), PRKN (24/2809, 0·9%), and SNCA (2/2809, <0·1%); two individuals were dual GBA1-LRRK2 carriers. Variant spectra for GBA1 and LRRK2 are shown in Figures 3 and 4, respectively. In controls, pathogenic or risk variants were identified in 1·8% (30/1654), exclusively involving GBA1.

Genetic variants in individuals with Parkinson’s disease of Central Asian ancestry (CAS) were identified in 8·2% (105/1284), primarily attributable to PD risk-associated variants in GBA1 (75/1284, 5·8%) and LRRK2 (23/1284, 1·8%) (Figures 3 and 4). Causal variants were rare and included variants in LRRK2 and RAB32 (<0·1% each, respectively), as well as PRKN variants identified in seven individuals (0·5%). Two individuals carried variants in more than one PD-associated gene (PRKN-GBA1 and LRRK2 risk-GBA1). In controls, pathogenic or risk variants were detected in 3·0% (44/1480), including GBA1 variants (n=29), PD risk-associated LRRK2 variants (n=14), and a single biallelic PRKN carrier.

In East Asian ancestry (EAS), 17·6% (840/4773) of PD individuals carried causative or risk variants, the majority harbouring one or both of the two LRRK2 risk variants p.R1628P and p.G2385R, identified in a total of 601 individuals (12·6%). GBA1 variants were the next most frequent with 190 carriers (4·0%), with a broad and heterogeneous variant spectrum (Figure 3), including over 40 distinct variants. Pathogenic LRRK2 variants were rare (n=15, 0·3%; Figure 4); other causal variants were found in SNCA (n=4, <0·1%), VPS35 (n=5, 0·1%), PINK1 (n=17, 0·4%), and PRKN (n=17, 0·4%). All biallelic PINK1 variant carriers harboured the same p.L347P variant in the homozygous state, while this variant was also observed in the heterozygous state in PD (13/4773, 0·3%) and controls (6/2809, 0·2%). 25 individuals with PD were dual carriers, most frequently a LRRK2 risk variant in combination with GBA1. Similarly, among the 6·7% (187/2809) of controls identified as carriers of a PD-associated pathogenic or risk variant, LRRK2 risk variants were most frequent (n=170).

Genetic findings were identified in 9·9% (4054/40773) of individuals with PD of European ancestry (EUR) and across a range of different PD-associated genes. GBA1 variants accounted for the largest proportion, with 3431 carriers (8·4%). Although the established PD risk variants p.E365K and p.T408M were the most frequent, a broad spectrum of additional GBA1 variants was identified across the European population (Figure 3). Pathogenic LRRK2 variants were the second most frequent finding and identified in 1·1% (453/40773) individuals, encompassing a broad range of variants, including several predominantly or exclusively observed in this ancestry (Figure 4). LRRK2 risk variants were detected in a small subset (n=21, <0·1%). About 0·1% (49/40773) were dual carriers, most frequently involving GBA1 in combination with LRRK2. In addition to GBA1 and LRRK2, variants were detected in several additional PD-associated genes, including SNCA (n=31, <0·1%), VPS35 (n=8, <0·1%), and RAB32 (n=17, <0·1%). Among recessive genes, biallelic variants were most frequently identified in PRKN (126/40773, 0·3%), whereas PINK1 (n=5) and PARK7 (n=3) variants were rare (both <0·1%, respectively). In comparison, 5·0% (1333/26 753) of controls carried pathogenic or risk variants, most often involving GBA1, with fewer pathogenic LRRK2 or other variants.

We observed a causal or risk variant in 20·5% of individuals with Parkinson’s disease (26/127) of Finnish ancestry (FIN). One individual carried a biallelic PRKN with a co-occurring GBA1 variant, while the remaining individuals carried GBA1 variants only. Similarly, genetic findings were confined to GBA1 among controls (4/22, 18·2%).

In individuals with Parkinson’s disease of Middle Eastern ancestry (MDE), PD-associated causal and risk variants were identified in 11·7% (158/1347) and distributed across multiple PD-associated genes. Most frequent were variants in GBA1 (n=73, 5·4%) and LRRK2 (n=59, 4·4%), with the LRRK2 variant spectrum largely driven by p.G2019S (Figure 4), whereas GBA1 showed a more heterogeneous spectrum (Figure 3). Additional findings included variants in PRKN (n=17, 1·3%), PINK1 (n=11, 0·8%), RAB32 (n=4, 0·3%), and SNCA (n=1, <0·1%). Seven individuals were dual carriers of GBA1 and LRRK2 variants. Among controls, pathogenic or risk variants were detected in 1·9% (25/1339), most frequently involving GBA1.

Causal or risk variants were identified in 6·1% (49/807) of individuals with Parkinson’s disease of South Asian ancestry (SAS). GBA1 variants were the most frequent findings (n=39, 4·8%), with p.L483R being the most frequent variant (Figure 3). Variants in recessive PD genes were also detected, including PRKN (n=5, 0·6%) and PINK1 (n=4, 0·5%), while pathogenic variants in LRRK2 and VPS35 were each identified in a single individual only (0·1% each, respectively). Among controls (4/440; 0·9%), only GBA1 variants were identified.

Across all ancestries, we identified 1,182 individuals with PD carrying single heterozygous pathogenic variants in recessive genes (1182/58559, 2·0%), most frequently individuals carrying single heterozygous variants in PRKN (n=1064), followed by PINK1 (n=40), VPS13C (n=23), ATP13A2 (n=22), and FBXO7 (n=21) whereas variants in JAM2, PARK7, and SYNJ1 were rarer (n<20). A substantial proportion of these, 41·7%, had an AAO ≤50 years (377/904; AAO missing for n=278). The overall proportion of individuals with an AAO ≤50 years among the entire PD group included in this study was 26·4% (11814/44768; AAO missing for n=13791). In comparison, the frequency of single heterozygous pathogenic variants in recessive genes in controls was 1·4% (583/41224).

Eleven individuals carried two heterozygous PRKN variants but had an AAO >60 years or the AAO was unavailable, so we did not interpret them as likely compound heterozygous without further confirmation.

Discussion

In this multi-ancestry investigation, we delineated shared and ancestry-specific patterns of genetic risk and causation across established PD genes, extending prior work that has largely focused on European ancestry and smaller, ancestry-restricted cohorts. Our work represents the largest ancestry-informed assessments to date, including about 100,000 individuals with ~30% of non-European and non-Ashkenazi Jewish ancestry individuals, offering insights into the ancestry-specific genetic architecture of PD. This characterization of genetic PD across diverse ancestries is important as multiple ongoing or planned clinical trials target proteins encoded by PD-linked genes.^7,20 Ensuring these advances are inclusive requires a detailed understanding of genetic variation beyond European ancestry.²¹

GBA1 risk variants were most frequent overall and identified across all populations, albeit with distinct ancestry-specific variant spectra: p.N409S was the most common variant in AJ, whereas p.E365K and p.T408M were more frequent in EUR, FIN, and CAS, but rare in AFR, AAC, and EAS. The variant spectrum in EAS was broad, but p.L483P appeared more frequent than in other populations. Similarly, we newly highlight p.L483R as a recurrent variant in South Asian individuals. The intronic variant rs3115534-G was the most frequently observed in AAC and AFR, as previously reported.¹⁰ This variant represents a clear example of an ancestry-specific GBA1 risk allele that would have remained undetected in studies restricted to European-ancestry populations, underscoring the necessity of ancestry-diverse genetic investigations. Collectively, these findings emphasize the importance of GBA1 as a globally relevant therapeutic target, an important observation given that most trials targeting glucocerebrosidase are conducted in countries with a high representation of European or Ashkenazi Jewish populations.⁴

LRRK2 variants also showed ancestral variability. p.G2019S was detected across multiple ancestries, with the highest frequencies in MDE (partially reflecting North African Berbers), AMR, and AJ.^22,23 While p.G2019S was also the most frequent LRRK2 variant in EUR, the mutational spectrum was notably broader and included p.R1441C, p.R1441G, and p.L1795F with suggested founder effects in European sub-populations.^24-27 In contrast, p.G2019S was largely absent from Asian populations, where the risk variants p.R1628P and p.G2385R were most frequent,²⁸ especially in EAS. The most frequent pathogenic LRRK2 variant identified in the EAS population was p.R1067Q, only recently proven to be disease-relevant and shown to be enriched in East Asians.¹³ Our data newly document the first pathogenic LRRK2 variant carriers of African ancestry, supporting the relevance of LRRK2-mediated PD in populations that have been largely absent from genetic and trial efforts to date. Together, these observations illustrate that, while LRRK2 is a globally relevant therapeutic target, ancestry-specific studies are essential to capture distinct variant spectra. LRRK2 variants differ in their impact on kinase activity,²⁹ which may inform clinical trial design and therapeutic targeting. Investigating LRRK2 variants that influence kinase activity and may affect eligibility for, or response to, emerging LRRK2 kinase-directed therapies is therefore important globally, particularly given the ongoing LRRK2-targeted trials.

SNCA variants were overall rare but identified across multiple ancestries, in line with previous reports.³⁰ While SNCA multiplications were identified in several populations, missense variants (p.A53T and p.G51D) were predominantly observed in EUR. Other known autosomal dominant forms of PD caused by VPS35 p.D620N or RAB32 p.S71R were rare. RAB32 variants were only detected in EUR and MDE individuals, more frequent in the latter, while VPS35 variant carriers were observed across three different populations, interestingly with an apparent enrichment in Asian populations (EAS and SAS) compared with Europeans, an ancestry pattern that has not previously been described at this scale. Given the limited sample size of some populations, especially with sequencing data, these findings need cautious interpretation. Nonetheless, detecting VPS35 and RAB32 variants across multiple ancestries underscores the importance of continued investigation in larger, diverse cohorts.

Finally, investigating recessive genes confirms PRKN as the most frequently implicated cause of recessive PD across nearly all ancestries and indicates a wider-than-appreciated distribution of PINK1 variants, underscoring the global relevance of PINK1/Parkin pathway-based approaches. In Asian populations (EAS, CAS, and SAS), PRKN variants were more frequent than pathogenic LRRK2 variants. Importantly, identifying PRKN carriers across multiple ancestries reinforces its global relevance, as preclinical efforts increasingly explore therapeutic strategies targeting Parkin or its associated pathways.⁷ While PINK1 variants were rare, we identified a notable proportion of MDE, EAS, and SAS carriers. All EAS PINK1 carriers harbored p.L347P, a variant known to be enriched in East Asian populations.³¹ We further identified numerous individuals carrying single heterozygous pathogenic variants in recessive PD genes. While these are not causative, many carriers had an early AAO ≤50 years, suggesting a second pathogenic variant, potentially a complex structural variant not captured by genotyping or short-read sequencing.

This global analysis underscores the clinical and translational importance of extending PD genetics research beyond European ancestry. Ancestry-informed genetic studies improve diagnostic accuracy, variant interpretation, and genotype-phenotype correlations, with direct implications for prognosis and genetic counselling. Our findings further show that variants predominantly identified in European populations often contribute only modestly in other ancestries, suggesting additional ancestry-specific variants and highlighting that our current understanding of the genetic architecture of PD remains incomplete. From a translational perspective, such diversity is essential for equitable clinical trial design, accurate identification of trial-eligible individuals, and the global applicability of emerging genetically targeted therapies. Together, these data provide a robust foundation for future in-depth ancestry-specific analyses and genetically informed therapeutic development efforts in PD.

We consider our efforts complementary to ongoing studies investigating common genetic risk through GWAS in diverse populations ^10-12,32, which continue to expand our understanding of PD susceptibility across ancestries. In this context, we focused specifically on established causative variants and risk-associated variants in genes involved in pathways targeted by ongoing clinical trials, while acknowledging that other well-established PD risk loci also contribute substantially to disease susceptibility, e.g., common risk variants at the SNCA locus and the MAPT H1 haplotype. Although rare causal variants in SNCA were included in our analyses because of their established direct role in disease pathogenesis, common GWAS-associated risk variants at the SNCA and MAPT loci were not included, as they are currently not used for genetically stratified clinical trial recruitment or therapeutic targeting in the same way as GBA1 and LRRK2 risk variants; however, they are increasingly explored in large-scale GWAS and dedicated haplotype analyses across diverse populations ³³.

Our study has limitations. It focused on variants in known PD genes, primarily identified in European populations. This specific focus and the predominance of European-ancestry data in resources like ClinVar may bias variant interpretation in non-European groups. While our study allowed for a large-scale, clinically relevant assessment of those known variants, it is not designed to lead to the discovery of novel PD genes and variants that may be more relevant in other populations. Systematically studying large, well-powered non-European ancestry cohorts using sequencing data will be important to capture the full global spectrum of novel variation. Another important area will be investigating variants of uncertain significance in PD genes. Some variants may be pathogenic in specific populations but are currently classified as of uncertain significance due to limited ancestry-specific reference data and low statistical power, whereas other variants may be overinterpreted in populations for which adequate reference data are lacking. Ancestry-aware (re)classification using variant frequency, functional, and segregation data will be key for establishing pathogenicity, improving genetic counselling and enabling trial access.^13,27

Some populations were of limited sample size. This limitation may inflate the relative contribution of known genes in well-powered populations and is particularly relevant for groups with limited WGS data, since some variants are not captured by genotyping (e.g., RAB32 p.S71R and select GBA1 variants).

The detection of compound-heterozygous variants was limited by the absence of phased data, preventing us from determining whether multiple variants were located on the same or opposite parental alleles. We conservatively considered carriers of two heterozygous pathogenic variants in recessive genes with an AAO ≤60 years as likely compound heterozygous, but validation is needed. Although genetically inferred ancestry provides a more robust framework than geographic labels, the broad ancestry categories still encompass substantial heterogeneity and may reflect cohort-specific recruitment patterns. Nonetheless, we believe this approach represents the most transparent and scalable strategy currently feasible for large, harmonised cross-ancestry analyses.

Lastly, findings were generated in a research setting; while we employed quality control measures and WGS validation where possible, further confirmation of identified variants in certified diagnostic laboratories is required before clinical implementation.

In conclusion, this study represents the largest global assessment of the genetic spectrum of PD, offering crucial insights into the population-specific genetic architecture of PD and underscoring the critical need to expand PD genetics research beyond European ancestry. Ancestry-informed analyses enhance diagnostic precision and risk interpretation and are essential to ensure equitable access to genetically stratified clinical trials and emerging precision therapies. Addressing current gaps in data diversity and global genetic representation is not optional but fundamental to translating genetic discovery into effective and inclusive genetically informed therapies in PD on a global scale.

Supplementary Material

NIHMS2198437-supplement-1.pdf^{(6.4MB, pdf)}

NIHMS2198437-supplement-2.xlsx^{(94.9KB, xlsx)}

Research in context.

Evidence before this study

The identification of rare causal and more common risk variants has dramatically advanced our understanding of Parkinson’s disease (PD). We searched PubMed from January 2020 to December 2025 for Articles reporting genetic screening studies in Parkinson’s disease, available in English, using search terms “Gene*”, “Parkinson’s disease”, and “screen”. The majority of genetic data has been generated in PD cohorts of European ancestry. The sample sets of two recent comprehensive, large-scale genetic screening studies, the PD GENEration North America and the Rostock International Parkinson’s disease (ROPAD) study, comprised 86% (6978/8159) and 92% (11579/12580) of White participants, respectively. Moreover, a global survey of monogenic PD, published in 2023, reported that 91% of individuals with identified genetic causal or risk variants in PD-associated genes, including individuals with Parkinson’s disease and unaffected individuals, were White, underscoring the major gap in ancestral representation in PD genetics research. Although ancestry-specific variants and founder effects have been reported for some PD-associated genes such as GBA1, LRRK2, and PINK1, comprehensive and ancestry-diverse studies are missing, thereby constraining the applicability of genetic diagnostics and the development of equitable and inclusive genetically and ancestry-informed therapeutic strategies.

Added value of this study

This study, conducted within the Global Parkinson’s Genetics Program (GP2), represents the largest multi-ancestry genetic investigation of PD to date, including ~100,000 individuals from around the world, with ~29,000 participants of non-European and non-Ashkenazi Jewish ancestry (representing ~29% of the study cohort). This study includes the largest systematically curated genetic resource currently available for multiple populations historically underrepresented in genetic studies, for whom sample collection and data curation are logistically challenging and genetic reference data remain sparse. Even where absolute sample sizes are smaller than in European cohorts, these data provide essential information for ancestry-specific variant interpretation and disease biology. By exploring causal variants in established PD-linked genes and PD risk-associated variants across 11 diverse genetically defined ancestry groups, this study provides a comprehensive view of the genetic architecture of PD. This work highlights both shared and ancestry-specific contributions. Variability in GBA1 is a globally relevant genetic contributor to PD risk, with variant carriers identified across all investigated populations, reinforcing the relevance of GBA1-targeted therapeutic approaches across ancestries. Similarly, LRRK2 variant carriers, currently being recruited into LRRK2-targeted clinical trials, and individuals with PRKN causal variants, for whom PRKN-targeted approaches are being explored in experimental studies, were identified across multiple ancestral groups, supporting the broad applicability of genetically targeted approaches. At the same time, substantial ancestry-specific differences underscore the limited transferability of current PD gene and variant findings, which are largely derived from European-ancestry datasets. This observation underscores the need to account for ancestry differences when evaluating the pathogenicity and clinical implications of PD-associated variants. Our findings also illustrate how variants identified in populations underrepresented in genetic studies often remain of uncertain significance due to limited comparative data.

Implications of all the available evidence

This study establishes a foundation for future ancestry-specific analyses, functional studies, and clinically actionable genetic research. As genetically informed therapeutic approaches are integrated into PD research and clinical trials, it is critical to ensure that emerging therapies are applicable and accessible to individuals of all ancestral backgrounds. Current genetically stratified trials predominantly enrol participants of European or Ashkenazi Jewish ancestry and trial sites are predominantly located across Europe and North America, raising concerns about equity, generalisability, and global clinical impact. This study underscores the importance of inclusive genetic efforts such as GP2 to close these existing gaps, improve diagnostic equity, inform trial design, and enable equitable and effective implementation of genetically informed therapeutics in Parkinson’s disease worldwide.

Acknowledgments

This project was supported by the Global Parkinson’s Genetics Program (GP2; https://gp2.org). GP2 is funded by the Aligning Science Across Parkinson’s (ASAP) initiative and implemented by The Michael J. Fox Foundation for Parkinson’s Research (MJFF). For a complete list of GP2 members see doi.org/10.5281/zenodo.7904831.

This research was supported in part by the Intramural Research Program of the National Institutes of Health (NIH). The contributions of the NIH authors are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.

We thank Yuliia Kanana for her valuable support in processing and handling the samples used in this study. We are grateful for the important support from all study participants and their families as well as for contributions of all collaborators, brain banks and biobanks.

Declaration of Interests

LML, SoB, MLD, PH, SJ, SUR, LS, CS, SS, EJS, AHT, ZT, NZ, KL, and SBC declare no relevant conflict of interests. LML received faculty honoraria from the Movement Disorder Society in the past 12 months, unrelated to this manuscript. ZHF is supported by the Aligning Science Across Parkinson’s (ASAP) Global Parkinson’s Genetics Program (GP2) and receives GP2 salary support from The Michael J. Fox Foundation for Parkinson’s Research. ZHF, MBM, NK, ShB, HI, LJ, MJK, HLL, KSL, DV, and MAN’s participation in this project was part of a competitive contract awarded to DataTecnica LLC by the National Institutes of Health to support open science research. MAN also currently serves on the scientific advisory board for Character Bio Inc plus is a scientific founder at Neuron23 Inc and owns stock. KAB and MT are supported by the Aligning Science Across Parkinson’s (ASAP) Global Parkinson’s Genetics Program (GP2). PH serves as an advisor to Alector Inc., the Global Parkinson’s Genetics Consortium (Michael J. Fox Foundation), LSP Advisory B.V., and Neuro.VC. HH and RK received essential funding from The Wellcome Trust, The MRC, The MSA Trust, The National Institute for Health Research University College London Hospitals Biomedical Research Centre NIHR-BRC), The Michael J Fox Foundation (MJFF), The Fidelity Trust, Rosetrees Trust, The Dolby Family fund, Alzheimer's Research UK (ARUK), MSA Coalition, The Guarantors of Brain, Cerebral Palsy Alliance, FARA, EAN and the NIH NeuroBioBank, Queen Square BrainBank, and The MRC Brainbank Network. JJ was supported by the Michael J. Fox Foundation (Data Community Innovators Program). KRK is supported by the Ainsworth 4 Foundation and the Medical Research Future Fund. SYL received consultancies and grants from the Michael J. Fox Foundation for Parkinson’s research (MJFF) and the Aligning Science Across Parkinson’s (ASAP) Global Parkinson’s Genetics Program (GP2). Unrelated to this manuscript, SYL received honoraria for participating as a Member of the Neurotorium Editorial Board and honoraria for lecturing/teaching from the International Parkinson and Movement Disorder Society (MDS) and Medtronic. SYL also received stipends from the MDS as Chair of the Asian-Oceanian Section, and npj PD as Associate Editor. NEM receives NIH funding (1K08NS131581). NEM is supported by the Aligning Science Across Parkinson’s (ASAP) Global Parkinson’s Genetics Program (GP2) and is member of the steering committee of the PD GENEration study for which he receives an honorarium from the Parkinson’s Foundation. WMYM is the founding member and Chair of the AfrAbia Parkinson’s Disease Genomic Consortium (AfrAbia PD-GC), which is supported by funding from the Global Parkinson’s Genetics Program (GP2). AJN reports grants from Parkinson's UK, Barts Charity, Cure Parkinson’s, National Institute for Health and Care Research, Innovate UK, the Medical College of Saint Bartholomew’s Hospital Trust, Alchemab, Aligning Science Across Parkinson’s Global Parkinson’s Genetics Program (ASAP-GP2) and the Michael J Fox Foundation and consultancy and personal fees from AstraZeneca, AbbVie, Profile, Bial, Charco Neurotech, Alchemab, Sosei Heptares, Umedeor and Britannia. AJN is an Associate Editor for the Journal of Parkinson’s Disease. RO has received travel grants from the Movement Disorder Society in the past 12 months, unrelated to this manuscript, and received research grants and travel support to attend annual GP2 meetings from the Aligning Science Across Parkinson’s (ASAP) Global Parkinson’s Genetics Program (GP2). NUO reports funding from MJFF and UK NIHR institutional grant funding, both for PD research. AHT receives support from the Michael J Fox Foundation and the Global Parkinson Genetic Program (GP2). Unrelated to this manuscript, AHT received speaker honoraria from International Parkinson and Movement Disorders, Eisai and Orion Pharma and reports consultancies from Elsevier as Section Editor for Parkinsonism and Related Disorders. JT is supported by the German Research Foundation (DFG), Aligning Science Across Parkinson’s (ASAP) Global Parkinson’s Genetics Program (GP2), and is a consultant for Acurex. BT acknowledges support from the Global Parkinson’s Genetics Program (GP2). EMV is supported by the Aligning Science Across Parkinson’s (ASAP) Global Parkinson’s Genetics Program (GP2). Unrelated to this manuscript, she received research grants from Telethon Foundation Italy (GGP20070), the Italian Ministry of Health (Ricerca Finalizzata RF-2019-12369368 and ERA-NET Neuron NDCil project EUR002), the Italian Ministry of University and Research (GENERA, MNESYS) and the Silverstein Foundation. KL received grants from the Dystonia Medical Research Foundation and German Research Foundation (DFG), unrelated to this manuscript. CB is an employee of the Coalition for Aligning Science (CAS). ABS is the lead investigator for a grant from the Michael J Fox Foundation for Parkinson’s Research and has a contract for work on the Global Parkinson’s Genetics Program (GP2). Unrelated to this manuscript, ABS is named as an inventor on patents for a diagnostic for stroke and for molecular testing for C9orf72 repeats; ABS is on the scientific advisory board of the Lewy Body Disease Association and for Cajal Neuroscience (both unpaid positions); and ABS received an honorarium for speaking at the World Laureates Association. HRM reports grant support from Parkinson’s UK, Cure Parkinson’s Trust, PSP Association, Medical Research Council, and the Michael J Fox Foundation. Unrelated to this manuscript, HRM is a co-applicant on a patent application related to C9orf72 - Method for diagnosing a neurodegenerative disease (PCT/GB2012/052140) and received honoraria from the Movement Disorder Society. CK received grants from the Michael J. Fox Foundation for Parkinson’s Research and the Aligning Science Across Parkinson’s Initiative. Unrelated to this manuscript, CK received grants from the German Research Foundation and speakers’ honoraria from Bial. CK has royalties at Oxford University Press and Springer Nature and serves as a medical advisor to Centogene and Biogen.

Footnotes

Data and Code Availability

Data used in the preparation of this article were obtained from the Global Parkinson’s Genetics Program (GP2; https://gp2.org). Specifically, we used Tier 2 data from GP2 release 11 (https://doi.org/10.5281/zenodo.17753486). Tier 1 data can be accessed by completing a form on the Accelerating Medicines Partnership in Parkinson’s Disease (AMP^®-PD) website (https://amp-pd.org/register-for-amp-pd). Tier 2 data access requires approval and a Data Use Agreement signed by your institution. Qualified researchers are encouraged to apply for direct access to the data through AMP PD.

All code generated for this article, and the identifiers for all software programs and packages used, are available on GitHub (https://github.com/GP2code/GP2-global-genetic-variant-landscape) and were given a persistent identifier via Zenodo (DOI 10.5281/zenodo.15699539). A detailed list of all identified variant carriers, including corresponding anonymised GP2-IDs, variant details as well as basic demographic and clinical characteristics are available to qualified researchers and upon reasonable request from the corresponding author.

References

1.Lim S-Y & Klein C. Parkinson’s Disease is Predominantly a Genetic Disease. J. Parkinsons. Dis. 14, 467–482 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Leonard HL & Global Parkinson’s Genetics Program (GP2). Novel Parkinson’s disease genetic risk factors within and across European populations. medRxiv (2025) doi: 10.1101/2025.03.14.24319455. [DOI] [Google Scholar]
3.Lange LM et al. Nomenclature of Genetic Movement Disorders: Recommendations of the International Parkinson and Movement Disorder Society Task Force - an update. Mov. Disord. 37, 905–935 (2022). [DOI] [PubMed] [Google Scholar]
4.Lim S-Y et al. Uncovering the genetic basis of Parkinson’s disease globally: from discoveries to the clinic. Lancet Neurol. 23, 1267–1280 (2024). [DOI] [PubMed] [Google Scholar]
5.Junker J et al. Team science approaches to unravel monogenic Parkinson’s disease on a global scale. Mov. Disord. 39, 1868–1873 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Vollstedt E-J et al. Embracing monogenic Parkinson’s disease: The MJFF Global Genetic PD cohort. Mov. Disord. 38, 286–303 (2023). [DOI] [PubMed] [Google Scholar]
7.Senkevich K, Rudakou U & Gan-Or Z. New therapeutic approaches to Parkinson’s disease targeting GBA, LRRK2 and Parkin. Neuropharmacology 202, 108822 (2022). [DOI] [PubMed] [Google Scholar]
8.Global Parkinson’s Genetics Program. GP2: The Global Parkinson’s Genetics Program. Mov. Disord. 36, 842–851 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Lange LM et al. Elucidating causative gene variants in hereditary Parkinson’s disease in the Global Parkinson's Genetics Program (GP2). NPJ Parkinsons Dis 9, 100 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Rizig M et al. Identification of genetic risk loci and causal insights associated with Parkinson’s disease in African and African admixed populations: a genome-wide association study. Lancet Neurol. 22, 1015–1025 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Leal TP et al. Genotype-phenotype association study conducted on LARGE-PD reveals novel loci associated with Parkinson’s Disease. medRxiv (2025) doi: 10.1101/2025.07.18.25331793. [DOI] [Google Scholar]
12.Kishore A et al. Deciphering the genetic architecture of Parkinson’s disease in India. medRxiv (2025) doi: 10.1101/2025.02.17.25322132. [DOI] [Google Scholar]
13.Lim S-Y et al. Clinical and functional evidence for the pathogenicity of the LRRK2 p.Arg1067Gln variant. NPJ Parkinsons Dis. 11, 34 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Velez-Pardo C et al. The distribution and risk effect of GBA variants in a large cohort of PD patients from Colombia and Peru. Parkinsonism Relat. Disord. 63, 204–208 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Cook L et al. Parkinson’s disease variant detection and disclosure: PD GENEration, a North American study. Brain 147, 2668–2679 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Westenberger A et al. Relevance of genetic testing in the gene-targeted trial era: the Rostock Parkinson’s disease study. Brain 147, 2652–2667 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Vitale D et al. GenoTools: An open-source Python package for efficient genotype data quality control and analysis. G3 (Bethesda) (2024) doi: 10.1093/g3journal/jkae268. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Kuznetsov N et al. CNV-Finder: Streamlining copy number variation discovery. bioRxivorg (2024) doi: 10.1101/2024.11.22.624040. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Lesage S et al. Characterization of recessive Parkinson’s disease in a large multicenter study. Ann. Neurol. 88, 843–850 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.McFarthing K et al. Parkinson’s disease drug therapies in the clinical trial pipeline: 2023 update. J. Parkinsons. Dis. 13, 427–439 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Schumacher-Schuh AF et al. Underrepresented populations in Parkinson’s genetics research: Current landscape and future directions. Mov. Disord. 37, 1593–1604 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.El Otmani H, Daghi M, Tahiri Jouti N & Lesage S. An overview of the worldwide distribution of LRRK2 mutations in Parkinson’s disease. Neurodegener. Dis. Manag. 13, 335–350 (2023). [DOI] [PubMed] [Google Scholar]
23.Lesage S et al. LRRK2 haplotype analyses in European and North African families with Parkinson disease: a common founder for the G2019S mutation dating from the 13th century. Am. J. Hum. Genet. 77, 330–332 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Gorostidi A, Ruiz-Martínez J, Lopez de Munain A, Alzualde A & Martí Massó JF. LRRK2 G2019S and R1441G mutations associated with Parkinson’s disease are common in the Basque Country, but relative prevalence is determined by ethnicity. Neurogenetics 10, 157–159 (2009). [DOI] [PubMed] [Google Scholar]
25.Mata IF et al. LRRK2 R1441G in Spanish patients with Parkinson’s disease. Neurosci. Lett. 382, 309–311 (2005). [DOI] [PubMed] [Google Scholar]
26.Nuytemans K et al. Founder mutation p.R1441C in the leucine-rich repeat kinase 2 gene in Belgian Parkinson’s disease patients. Eur. J. Hum. Genet. 16, 471–479 (2008). [DOI] [PubMed] [Google Scholar]
27.Lange LM et al. The LRRK2 p.L1795F variant causes Parkinson’s disease in the European population. NPJ Parkinsons Dis. 11, 58 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Lim S-Y et al. Parkinson’s disease in the Western Pacific Region. Lancet Neurol. 18, 865–879 (2019). [DOI] [PubMed] [Google Scholar]
29.Kalogeropulou AF et al. Impact of 100 LRRK2 variants linked to Parkinson’s disease on kinase activity and microtubule binding. Biochem. J. 479, 1759–1783 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Trinh J et al. Genotype-phenotype relations for the Parkinson’s disease genes SNCA, LRRK2, VPS35: MDSGene systematic review. Mov. Disord. 33, 1857–1870 (2018). [DOI] [PubMed] [Google Scholar]
31.Patel SG et al. Potential PINK1 founder effect in Polynesia causing early-onset Parkinson’s disease. Mov. Disord. 36, 2199–2200 (2021). [DOI] [PubMed] [Google Scholar]
32.Kim JJ et al. Multi-ancestry genome-wide association meta-analysis of Parkinson’s disease. Nat. Genet. 56, 27–36 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Reyes-Pérez P et al. Exploring MAPT-containing H1 and H2 haplotypes in Parkinson’s disease across diverse populations. medRxiv (2025) doi: 10.1101/2025.10.17.25337033. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS2198437-supplement-1.pdf^{(6.4MB, pdf)}

NIHMS2198437-supplement-2.xlsx^{(94.9KB, xlsx)}

[R1] 1.Lim S-Y & Klein C. Parkinson’s Disease is Predominantly a Genetic Disease. J. Parkinsons. Dis. 14, 467–482 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] 2.Leonard HL & Global Parkinson’s Genetics Program (GP2). Novel Parkinson’s disease genetic risk factors within and across European populations. medRxiv (2025) doi: 10.1101/2025.03.14.24319455. [DOI] [Google Scholar]

[R3] 3.Lange LM et al. Nomenclature of Genetic Movement Disorders: Recommendations of the International Parkinson and Movement Disorder Society Task Force - an update. Mov. Disord. 37, 905–935 (2022). [DOI] [PubMed] [Google Scholar]

[R4] 4.Lim S-Y et al. Uncovering the genetic basis of Parkinson’s disease globally: from discoveries to the clinic. Lancet Neurol. 23, 1267–1280 (2024). [DOI] [PubMed] [Google Scholar]

[R5] 5.Junker J et al. Team science approaches to unravel monogenic Parkinson’s disease on a global scale. Mov. Disord. 39, 1868–1873 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Vollstedt E-J et al. Embracing monogenic Parkinson’s disease: The MJFF Global Genetic PD cohort. Mov. Disord. 38, 286–303 (2023). [DOI] [PubMed] [Google Scholar]

[R7] 7.Senkevich K, Rudakou U & Gan-Or Z. New therapeutic approaches to Parkinson’s disease targeting GBA, LRRK2 and Parkin. Neuropharmacology 202, 108822 (2022). [DOI] [PubMed] [Google Scholar]

[R8] 8.Global Parkinson’s Genetics Program. GP2: The Global Parkinson’s Genetics Program. Mov. Disord. 36, 842–851 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Lange LM et al. Elucidating causative gene variants in hereditary Parkinson’s disease in the Global Parkinson's Genetics Program (GP2). NPJ Parkinsons Dis 9, 100 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Rizig M et al. Identification of genetic risk loci and causal insights associated with Parkinson’s disease in African and African admixed populations: a genome-wide association study. Lancet Neurol. 22, 1015–1025 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] 11.Leal TP et al. Genotype-phenotype association study conducted on LARGE-PD reveals novel loci associated with Parkinson’s Disease. medRxiv (2025) doi: 10.1101/2025.07.18.25331793. [DOI] [Google Scholar]

[R12] 12.Kishore A et al. Deciphering the genetic architecture of Parkinson’s disease in India. medRxiv (2025) doi: 10.1101/2025.02.17.25322132. [DOI] [Google Scholar]

[R13] 13.Lim S-Y et al. Clinical and functional evidence for the pathogenicity of the LRRK2 p.Arg1067Gln variant. NPJ Parkinsons Dis. 11, 34 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.Velez-Pardo C et al. The distribution and risk effect of GBA variants in a large cohort of PD patients from Colombia and Peru. Parkinsonism Relat. Disord. 63, 204–208 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R15] 15.Cook L et al. Parkinson’s disease variant detection and disclosure: PD GENEration, a North American study. Brain 147, 2668–2679 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Westenberger A et al. Relevance of genetic testing in the gene-targeted trial era: the Rostock Parkinson’s disease study. Brain 147, 2652–2667 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.Vitale D et al. GenoTools: An open-source Python package for efficient genotype data quality control and analysis. G3 (Bethesda) (2024) doi: 10.1093/g3journal/jkae268. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Kuznetsov N et al. CNV-Finder: Streamlining copy number variation discovery. bioRxivorg (2024) doi: 10.1101/2024.11.22.624040. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Lesage S et al. Characterization of recessive Parkinson’s disease in a large multicenter study. Ann. Neurol. 88, 843–850 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.McFarthing K et al. Parkinson’s disease drug therapies in the clinical trial pipeline: 2023 update. J. Parkinsons. Dis. 13, 427–439 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] 21.Schumacher-Schuh AF et al. Underrepresented populations in Parkinson’s genetics research: Current landscape and future directions. Mov. Disord. 37, 1593–1604 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.El Otmani H, Daghi M, Tahiri Jouti N & Lesage S. An overview of the worldwide distribution of LRRK2 mutations in Parkinson’s disease. Neurodegener. Dis. Manag. 13, 335–350 (2023). [DOI] [PubMed] [Google Scholar]

[R23] 23.Lesage S et al. LRRK2 haplotype analyses in European and North African families with Parkinson disease: a common founder for the G2019S mutation dating from the 13th century. Am. J. Hum. Genet. 77, 330–332 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.Gorostidi A, Ruiz-Martínez J, Lopez de Munain A, Alzualde A & Martí Massó JF. LRRK2 G2019S and R1441G mutations associated with Parkinson’s disease are common in the Basque Country, but relative prevalence is determined by ethnicity. Neurogenetics 10, 157–159 (2009). [DOI] [PubMed] [Google Scholar]

[R25] 25.Mata IF et al. LRRK2 R1441G in Spanish patients with Parkinson’s disease. Neurosci. Lett. 382, 309–311 (2005). [DOI] [PubMed] [Google Scholar]

[R26] 26.Nuytemans K et al. Founder mutation p.R1441C in the leucine-rich repeat kinase 2 gene in Belgian Parkinson’s disease patients. Eur. J. Hum. Genet. 16, 471–479 (2008). [DOI] [PubMed] [Google Scholar]

[R27] 27.Lange LM et al. The LRRK2 p.L1795F variant causes Parkinson’s disease in the European population. NPJ Parkinsons Dis. 11, 58 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R28] 28.Lim S-Y et al. Parkinson’s disease in the Western Pacific Region. Lancet Neurol. 18, 865–879 (2019). [DOI] [PubMed] [Google Scholar]

[R29] 29.Kalogeropulou AF et al. Impact of 100 LRRK2 variants linked to Parkinson’s disease on kinase activity and microtubule binding. Biochem. J. 479, 1759–1783 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R30] 30.Trinh J et al. Genotype-phenotype relations for the Parkinson’s disease genes SNCA, LRRK2, VPS35: MDSGene systematic review. Mov. Disord. 33, 1857–1870 (2018). [DOI] [PubMed] [Google Scholar]

[R31] 31.Patel SG et al. Potential PINK1 founder effect in Polynesia causing early-onset Parkinson’s disease. Mov. Disord. 36, 2199–2200 (2021). [DOI] [PubMed] [Google Scholar]

[R32] 32.Kim JJ et al. Multi-ancestry genome-wide association meta-analysis of Parkinson’s disease. Nat. Genet. 56, 27–36 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R33] 33.Reyes-Pérez P et al. Exploring MAPT-containing H1 and H2 haplotypes in Parkinson’s disease across diverse populations. medRxiv (2025) doi: 10.1101/2025.10.17.25337033. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Parkinson’s disease genetics across diverse ancestries: an observational genetic study of causal and risk variants with translational implications

Lara M Lange, MD

Zih-Hua Fang, PhD

Mary B Makarious, PhD

Nicole Kuznetsov, MPS

Kajsa Atterling Brolin, PhD

Shannon Ballard, PhD

Prof Soraya Bardien, PhD

Maria Leila Doquenia, MD

Prof Peter Heutink, PhD

Prof Henry Houlden, MD

Hirotaka Iwaki, PhD

Simona Jasaityte, MSc

Lietsel Jones, MSc

Johanna Junker, MD

Rauan Kaiyrzhanov, MD

Mathew J Koretsky, BSc

Kishore R Kumar, PhD

Hampton L Leonard, MSc

Kristin S Levine, MSc

Prof Shen-Yang Lim, MD

Niccoló E Mencacci, MD

Prof Wael M Y Mohamed, PhD

Mike A Nalls, PhD

Prof Alastair J Noyce, PhD

Rajeev Ojha, MD

Prof Njideka U Okubadejo, MD

Shoaib ur Rehman, PhD

Laurel Screven, PhD

Chingiz Shashkin, MD

Sophia Sopromadze, MD

Eleanor J Stafford, MSc

Prof Ai Huey Tan, MD

Manuela Tan, PhD

Zaruhi Tavadyan, MD

Joanne Trinh, PhD

Bayasgalan Tserensodnom, MD

Prof Enza Maria Valente, PhD

Dan Vitale, MSc

Nazira Zharkinbekova, MD

Prof Katja Lohmann, PhD

Sara Bandres-Ciga, PhD

Cornelis Blauwendraat, PhD

Andrew B Singleton, PhD

Prof Huw R Morris, PhD

Prof Christine Klein, MD

Abstract

Background:

Methods:

Findings:

Interpretation:

Funding:

Introduction

Methods

Study design and participants

Figure 1. Study Workflow and Implications.

Procedures

Statistical analysis

Role of the funding source

Results

Table 1.

Table 2.

Figure 2. Global genetic spectrum of Parkinson's disease and locations of clinical trials targeting gene variant carriers.

Figure 3. Mutational spectrum and pathogenicity of GBA1 variants across ancestries.

Figure 4. Distribution of LRRK2 variants across protein domains and ancestries.

Discussion

Supplementary Material

Research in context.

Evidence before this study

Added value of this study

Implications of all the available evidence

Acknowledgments

Declaration of Interests

Footnotes

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK