Welcome to the new GEoPD Website

    A multi-centre study of genotype-phenotype correlation in Parkin-Parkinson’s disease.


    • Suzanne Lesage
    • Alexis Brice
    • Jean- Christophe Corvol

    Background: Parkinson’s disease (PD) is caused by a complex interplay between genetic and environmental factors. Genome-wide association studies have identified more than 90 independent risk variants across more than 80 genomic regions associated with Parkinson’s disease risk and progression. Although PD is commonly sporadic, monogenic causes are identified in 30% of familial and 3-5% of patients with sporadic PD.

    Mutations in Parkin is the most frequent cause of autosomal recessive, early onset, Parkinson’s disease. Although Parkin-related PD is characterized by an early onset of motor symptoms, typically before the age of 40, a “pure” motor disease, and a slower disease progression than sporadic PD, there is a wide heterogeneity of disease profiles across patients. Several types of mutations have been associated with Parkin-related PD (missense mutations, frameshifts, rearrangements) and may contribute to this variability. However, the age at onset can vary by up to 20 or 30 years between patients with the same mutations or within the same family, suggesting the presence of as of yet undiscovered environmental or genetic modifiers. Ethnicity may also play a role in genotype as shown by the increased frequency of Parkin mutations in Caucasians compared to Arab-Berbers.

    Research significance

    In sporadic PD, genome wide association studies have shed light on genetic variants associated with motor onset and/or disease progression. There are also studies assessing genetic modifiers of risk and age at onset of GBA and LRRK2 associated Parkinson’s disease. However, there is no information at present on the genomic modifiers of age of onset, symptoms and progression in patients with biallelic mutations in Parkin. We hope this study will assist with informing about these genomic modifiers.

    Study Aims

    1. Evaluate the impact of different types of Parkin mutations (truncations, rearrangements, missense m      utations) on Parkin-related PD clinical profiles (age of onset, motor and non-motor symptoms).

    2. Identify possible genetic modifiers of age of onset and disease progression in Parkin PD patients using genome wide association studies and whole exome analysis.

    Preliminary results

    We have currently analysed for correlation between the type of mutation and clinical severity in Parkin-PD patients available from our own cohort at the ICM, Paris. Preliminary results have identified that there is a difference in the age of onset and clinical severity in patients who possess the same mutation, suggestive of other genomic and environmental modifiers of phenotype.


    Patients confirmed to have Parkinson’s disease secondary to biallelic pathogenic Parkin variants on sequencing, with no other pathogenic mutations implicated on a PD gene panel, will be identified for inclusion in this study.

    Clinical information consisting of age of onset of Parkinson’s disease, duration of symptoms at time of clinical examination, gender, Hoehn and Yahr stage (on and predicted off state), MDS-UPDRS part 3 scores in on state and MMSE or MOCA scores will be determined. Demographic information consisting of current age, gender, ethnicity and country of origin will be included.

    Age of onset will be dichotomized into early and late onset and considered as a quantitative trait. Whole-genome genotyping will be performed using highly dense DNA microarrays, such as Illumina Infinium OmniExpress-24 kits. Analysis will be performed to understand if there is a correlation between the type of mutation with the age at onset of PD as well as clinical severity. Whole exome and whole genomic sequencing (where possible) will be performed on samples from these patients with extreme phenotypes, using ICM genotyping/sequencing platforms. Loci associated with the age of onset of PD in these patients will be identified through linkage and association analysis. We will further deep sequence any association loci that is identified.

     Timeline & Resource’ required

    Clinical information and genomic samples are required from collaborators. The collection of data from collaborators, preparation of the dataset with information from collaborating centers and the receipt of DNA will take around 6 -12 months. Genomic sequencing and analysis of the genomic data will take another 12 -18 months. The resources necessary for data-mining, genomic sequencing and analysis of genomic data will be covered by our site.

    Contributors to this project will be included in any resulting publications.

    Action steps

    Dr. Poornima Menon, Movement disorders fellow will help co-ordinate this study and can be contacted if there are any queries at: poornima.menon@icm-institute.org  

    Collaborating centres should have obtained consent from participating patients. Provide detailed clinical information, genomic information about the Parkin variant and 20µg of DNA for genotyping and sequencing.

    Assessing the role of variations in dementia genes on the risk to develop LBD ­in GBA-mutation carriers


    • Humanitas University (Milan, Italy): Stefano Duga, Letizia Straniero
    • Parkinson Institute of Milan: Gianni Pezzoli, Anna Zecchinelli
    • Policlinico Hospital (Milan, Italy): Alessio Di Fonzo

    Introduction and rationale

    Lewy body dementia (LBD) includes Parkinson’s disease dementia (PDD) and dementia with Lewy bodies (DLB), both characterized by abnormal deposition of alpha-synuclein in the brain. Several studies suggest that genes such as SNCA, APOE, and GBA are involved in LBD. In particular, heterozygous mutations in GBA are associated with an increased risk of LBD and Parkinson’s disease (PD); however, it is not fully understood why individuals with the same GBA variant eventually might develop PD or LBD.
    In this study, we will compare the burden of deleterious variants in dementia-related genes in LBD vs. PD patients carriers of a PD-associated GBA variant. The identification of additional genetic factors increasing the LBD risk could be helpful for early diagnosis and for the discovery of novel therapeutic strategies.
     Patients and Methods
    We plan to collect a large number of LBD patients (DLB and PDD) already screened or to be screened for GBA variants. A relatively large cohort of PD patient with a deleterious variant in GBA has already be gathered by our group. The recruited patients will be screened using an HaloPlex NGS custom panel focused on dementia-related genes. All identified variations will be validated by Sanger sequencing.
    Preliminary results
    We have already studied a cohort composed of 204 PD and 138 Italian LBD patients heterozygous for at least one PD-related GBA mutation. The first analyses suggest a disproportionate burden of deleterious variant in dementia patients.
    Timeline of the project
    We plan to enroll patient samples or genotype data from collaborating groups by December 2020. Additional data/samples received after the deadline will in any case incorporated into the study if received by the end of March 2021. 

    Contact: Should you be interested, do not hesitate to liaise with Stefano Duga and Alessio Di Fonzo.

    U1-splicing mutations in monogenic PD genes

    Investigators: Rejko Krüger, Ibrahim Boussaad
    Defective pre-mRNA processing is known to represent a common cause of human diseases, with ~30% of all mutations causing aberrant splicing1,2. In fact, a recent study showed that approximately 10% of pathogenic missense variants predicted to alter protein coding essentially disrupt splicing1. For Parkinson’s disease (PD) pathogenesis, the dysregulation of splicing as an alternative mechanism contributing to the neurodegenerative process was not systematically addressed3. In our recent study, we described a case of rare early onset PD that is caused by the c.192G>C mutation in PARK74. This mutation, that is located at the last base of exon 3, disrupts the binding motif of the spliceosome complex U1 and leads to skipping of exon 3 during splicing, which results in a non-translated, truncated mRNA. The U1 binding motif (splice site donor) at every 3’-exon/5’-intron-boundary consists of the last three bases of the exon and the first six bases of the intron. It is recognized by the U1 small nuclear RNA (U1 snRNA), which is part of the U1 complex. Mutations in the motif that abolish U1 binding lead to the loss of exon recognition and, hence, the exon is skipped during splicing, which results in truncated mRNA. We have successfully explored to strategies to rescue the pathologic exon skipping in our cellular model. First, we used a genetic approach by introducing a genetically modified U1 small nuclear RNA that matches the mutated patient sequence, and restores binding of the spliceosome. Furthermore, we have identified a combination of two small molecules, phenylbutric acid and RECTAS, that, when administered together, can rescue exon skipping and cellular phenotypes. Looking beyond our rare familial case of PD, we assessed the general prevalence of U1 splice site mutations in PD with a burden analysis using the whole exome data of the Parkinson Disease Genetics Sequencing Consortium (PDGSC). Comparing 2710 PD cases to 5713 controls, we observed a significant burden for genome-wide mutations in cases4.
    In this study, we aim to systematically address (i) the contribution of U1 splice site mutations in PD by analyzing genomic data, (ii) to identify new disease causing mutations in familial cases with no known mutations, and (iii) to assess the efficacy of our combinatorial compound treatment in patient-derived cellular models with different splice site mutations and to further screen for small molecule compounds that can rescue aberrant splicing.
    Material and Methods
    For the computational analysis of splice site variants we will need whole exome/genome Data of PD cases and ideally, but not necessarily, of controls. Using our in house burden analysis algorithm, we will analyze these data for the presence of variants with a high mutation score that are uniquely found in cases. If patient-derived cells are available, we will use them to validate the splice site mutation experimentally. If patient material is not available, we will use a reporter construct to clone the mutation and assess its effect on splicing in vitro. Patient-cells with validated U1 splice site mutations will be used to test the described combinatorial treatment. Analysis of mRNA splicing as well as the rescue of cellular phenotypes will be used to determine the success of the treatment. In order to broaden the panel of potential treatment options, we will also use cells with validated mutations to screen for other small molecules that rescue pathological exon skipping. The screening will be performed on our automated cell culture and screening platform. To date we can screen a library of over 2500 compounds of which approx. 60% are FDA approved.
    If you wish to find out more, then contact Rejko.


    1. Soemedi, R. et al. Pathogenic variants that alter protein code often disrupt splicing. Nat. Genet. (2017). doi:10.1038/ng.3837
    2. Lim, K. H., Ferraris, L., Filloux, M. E., Raphael, B. J. & Fairbrother, W. G. Using positional distribution to identify splicing elements and predict pre-mRNA processing defects in human genes. Proc. Natl. Acad. Sci. 108, 11093–11098 (2011).
    3. La Cognata, V., D’Agata, V., Cavalcanti, F. & Cavallaro, S. Splicing: is there an alternative contribution to Parkinson’s disease? Neurogenetics 16, 245–263 (2015).
    4. Boussaad, I. et al. A patient-based model of RNA mis-splicing uncovers treatment targets in Parkinson’s disease. Sci. Transl. Med. 12, (2020).


    Investigators: Katerina Markopoulou

    GEO-PD includes 20 sites from 14 countries and 4 continents who have committed to collect and share DNAs and clinical data longitudinally for 15 years, for 3,000 Parkinson's disease cases.  If you wish to be involved in this project, please contact Katerina.

    CNV calling on the GEoPD cases and controls

    Investigators: Rejko Krüger, Patrick May

    Parkinson's disease (PD) is the one of the most common neurodegenerative diseases world-wide. The underlying genetic causes are still not fully understood. Unlike in other complex brain disease it seems that copy number variation (CNV) in PD has not studied in large-scale except very recently for the Latino-American population (Sarihan et al. 2020, Movement Disorders). GEoPD is a clinical well-characterized new, multi-ethnical large PD cohort that is currently analysed for underlying genetic association with PD in general but also for associations with clinical parameters and sub-phenotypes.
    The CNV analysis of the GEoPD cohort will help to understand the genetic contribution to PD as well find new genetic associations to PD and sub-phenotypes. Are the CNVs enriched in PD overlapping with known PD loci? Are CNVs associated with age-of-onset and other clinical parameters or phenotypes?
    We will use mainly the NeuroChip data from COURAGE-PD but also other array types from GEoPD together with co-variates like sex, age-of-onset disease status, population stratification to perform CNV analysis. As analysis pipeline we will use the established CNV pipeline from the Broad-Institute and the Epi25 Collaborative, of which we are part of the analysis team, that was recently used for the largest CNV analysis in Epilepsy (Niestroj et al. 2020, Brain). Genotypes as B-allele frequency files from the original genotype. The analysis workflow is based on the well-established PennCNV (Wang et al. 2007) pipeline that will be used to create GC wave-adjusted intensity files. We will then perform intensity and CNV load quality control on the samples, filtering the calls, delineation of rare CNVs using common CNVs, annotation of CNVs and finally perform CNV burden analysis and CNV breakpoint level association.
    Combined with data on rare variants, as available from the NeuroChip or ongoing sequencing efforts, we would also analyse potential compound heterozygosity of CNVs with rare disease-causing variants and large deletions and insertions.
    Clinical data needed: We would need also Age at Onset, Sex and principal components as covariates for our analysis.
    The CNV analysis of the GEoPD cohort will help to understand the genetic contribution to PD as well find new genetic associations to PD and sub-phenotypes. Are the CNVs enriched in PD overlapping with known PD loci? Are CNVs associated with age-of-onset and other clinical parameters or phenotypes.

    If you wish to find out more, then contact Rejko.








    DBS Project Proposal
    Aim 1:  Identify participant sites with active DBS programs that can provide detailed phenotype and either genotype information or DNA for de-identified genotyping.
    Aim 2:  Identify genetic factors that contribute to disease progression and outcomes in DBS patients.
    Aim 3: Determine the significance of phenotypic and genetic factors in the choice of candidates for DBS surgery.
    To pursue these aims, we seek active participation of an international working group, willing to work towards data harmonization and collaborative pooled analyses, including genotyping/sequencing for which appropriate consent would need to be developed. This effort can be coordinated within the context of GP2.
    More specifically both retrospective and prospective studies are proposed.  In the retrospective study, a detailed phenotypic and genotypic characterization of patients who already have undergone DBS will be obtained from participating sites. The prospective study will include the longitudinal annual follow up of these established DBS cases as well as enroll new cases.
    To facilitate this effort, a study protocol is presented for discussion. Potential participants are asked to complete a survey, included below, to provide details about the information able to be gathered at their sites. 
     Study protocol
    1. Phenotypic characterization at study entry
    For any new or established case, the following phenotypic information is desired:
    • neuropsychological testing
    • age at disease onset
    • disease duration
    • disease subtype (tremor-predominant/akinetic-rigid/mixed)
    • Information on family history
    • H&Y stage
    • Time of emergence and duration of levodopa complications
    • LD/LEDD requirements
    • Orthostatic hypotension
    • Urinary incontinence
    • Co-morbidities
    • CNS imaging (CT and MRI)
    1. The choice of genotyping of established and prospective cases will be discussed:
    • gene panel     
    • candidate genes
    • gene chip
    • exome sequencing
    • whole genome sequencing
    1. DNA will be made available for additional deidentified genotyping and data harmonization
    2. DBS targets included in the study protocol are STN, GPi or other. The choice of target, laterality and choice of DBS manufacturer will be determined by the individual site and included in the analysis.
    3. The post-operative assessment should include the following at annual intervals at years 1-5:
    • Neuropsychological testing
    • H&Y stage
    • LD/LEDD requirements
    • presence of levodopa complications
    • frequency of DBS reprogramming

    Those who want to contribute to the project, please contact: M. Farrer and K. Markopoulou

    LRRK2 modifier project

    Investigators: Prof. Alexis Brice, Paris Brain Institute
    The study focuses on the hypothesis that genetic modifiers may affect the LRRK2 G2019S penetrance and age at onset by modulating gene expression. Recently, Trinh et al. identified a tagging SNP (rs2421947) in DNM3 for which the alternative allele is associated with an earlier age at onset in Tunisian G2019S carriers1.

    The aim of the present study is to elucidate the genetic contribution to age at onset variability by identifying the possible effect of additional modifiers in G2019S linked PD. The strategy we use to find out loci of interest includes genome-wide association analysis for unrelated G2019S carriers using age at onset as both a quantitative and a qualitative trait. We have already recruited 1292 of LRRK2 G2019S patients thanks to our collaborators and our local cohort from three different ethnical background.

    Our preliminary findings from genome-wide association analysis did not show any significant association with the DNM3 locus. However, our exploratory analysis on the North African cohort (N = 492 samples) identified two suggestive SNPs (rs1762792, p value = 8.70 10-8; rs1360821 p value = 1.00 10-7) by GWAS on chromosome 13q31.3 using the age at onset as a quantitative trait. Furthermore, when age at onset was dichotomised by the median age at onset the two suggestive SNPs (rs1762792, p value = 1.20 10-7; rs1360821, p value = 2.00 10-7) were confirmed. We finally used Kaplan–Meier analyses to compare how age of PD onset varied with rs1762792 genotypes in all 276 G2019S carriers analysed. The CC homozygotes had a median age at onset of 47 years; CT heterozygotes had a median age at onset of 52 years; TT homozygotes had a median age at onset of 57 years. The median age at onset of LRRK2 parkinsonism in rs1762792 CC homozygotes was 10 years younger than that of TT homozygotes suggesting a protective role of the wild type allele. These promising results were replicated in our European cohort.

    We now aim to reach increase the power of the study by collecting more LRRK2 G2019S patients (DNA samples or genotyping data) and extend our GWAS results. To detect genes whose expression is associated with the significant locus we will search for expression quantitative trait loci (e QTL).In order to identify additional but rare variants we plan to use WGS/WES in LRRK2 carriers, with extreme ages at onset. Again including a large number of cases would provide an opportunity for identifying and validating the role of rare variants.

    GCH-1 mutation and PD

    Investigator: Beomseok (BJ) Jeon from Seoul National University Hospital

    Background: A mutation in the GTP cyclohydrolase 1 (GCH-1) gene is a cause of Dopa-responsive dystonia (DRD). Previous studies showed that the some family members of DRD with GCH-1 mutation have parkinsonism with decreased DAT binding which is an evidence of nigral degeneration. A large-scale meta-analysis of genome-wide association data identified GCH-1 as one of the risk loci of PD. These findings suggested that the GCH-1 mutation may be a risk factor for nigral degeneration causing PD. However, our group recently published a case of PD patient with GCH-1 mutation showing excellent clinical course.1 Serial follow-up of dopamine transporter imaging (FP-CIT PET) of the patient suggested that GCH-1 mutation may have unmasked the subclinical pathology, thus conferring excellent clinical course at least apparently.

    Hypothesis: The PD patients with GCH-1 mutation would present with earlier-onset, benign clinical course with later development of motor complications and cognitive decline, requiring low levodopa dose, and relatively milder decrease in DAT binding. The clinical features of 44 patients of PD with GCH-1 genetic variant (13 with DAT imaging) from previously published studies matched the hypothesis.1 However, the hypothesis requires further support with larger number of samples.

    Method: We would like to collect clinical information of PD patients with the GCH-1 mutation. We would like to collect the age of onset, type of GCH-1 mutation, disease duration, Hoehn and Yahr stage, motor complications and cognitive decline, levodopa equivalent dose and dopamine transporter imaging. The detailed form is summarized in the attached case report form. *The information does not have to be complete.

    Analysis: We will summarize the clinical features of PD patients with GCH-1 mutation and see if they match with our hypothesis.

    Perspective: This study will support the hypothesis that GCH-1 mutation is not the cause of nigral degeneration but unmasks subclinical pathology.

    Latest Forum Posts

    • No posts to display.

    Who's online

    We have 283 guests and no members online