Literature searches and data extraction

The content of MDSGene is based on genetic as well as phenotypic and clinical data extracted from the relevant literature following systematic screens of different resources (Lill et al., 2016; Kasten et al., 2018). Eligible articles need to be written in English and published in peer-reviewed journals. They are identified following systematic PubMed searches based on standardized search terms comprising the name of the genes (including aliases) and the disease/syndrome of interest. Demographic, clinical and genetic data are extracted adhering to a standardized data extraction protocol. Diagnoses displayed in MDSGene follow the recent recommendations of the International Parkinson disease and Movement Disorder (MDS) Task Force of Genetic Nomenclature in Movement Disorders (Marras et al., 2016). Whenever necessary, mutations are remapped to the human genome build 19, and mutation identifiers are renamed according to the Human Genome Variation Society (HGVS) nomenclature. Mutation carriers are only included if clinical data has been provided and indicated that he/she is affected by a movement disorder.

Pathogenicity scoring

Potential pathogenicity of reported variants is classified as “possible”, “probable”, or “definite” based on the following criteria: i) co-segregation with disease in the reported pedigrees and/or the number of reported mutation carriers, ii) frequency in ~120,000 ethnically diverse individuals from the gnomAD (Genome Aggregation Database) browser (, iii) CADD (“Combined Annotation Dependent Depletion") score as an in-silico measure of deleteriousness of genetic variants (Kircher et al., 2014, Rentzsch et al., 2018), and iv) reported molecular evidence from in-vivo and/or in-vitro studies. Each evidence domain was divided into four categories each accumulating specific “points”, weighted by category (see Table 1). Evidence domains “co-segregation with disease” and “presence of mutation-specific positive functional data” received the strongest weights in the pathogenicity grading. Finally, points were summed across categories and pathogenicity was graded as follows: benign (<5 points), possibly pathogenic (5-9 points), probably pathogenic (10-14 points), definitely pathogenic (>14 points). Reported genetic variants that have been classified as benign using this scoring algorithm are not included in MDSGene.

Table 1. Pathogenicity scoring scheme implemented in MDSGene:

Evidence Segregation Frequency (gnomAD) In-silico prediction
(CADD score)
Functional studies
Least Only a single heterozygous patient
(0 points)
(0 points)
(0 points)
Only negative reports or absence of studies
(0 points)
Suggestive ≥1 biallelic patient for recessive or ≥2 single heterozygous patients
for dominant genes or 1 family
(i.e. ≥2 affected mutation carriers)
(2 points)
(1 point)
10 to <15
(1 point)
1 positive study
(2 points)
Strong 2 families
(3 points)
(2 points)
15 to 20
(3 points)
2 positive studies or null allele
(4 points)
Highest >2 families
or ≥1 de novo
(6 points)
(3 points)
(5 points)
>2 positive studies
(6 points)
LRRK2 variant screen to assess LRRK2 kinase pathway activity

The Leucine rich repeat kinase (LRRK2) is a large multidomain protein harboring a tandem catalytic ROC-COR GTPase and kinase domains as well as an N-terminal armadillo, ankyrin and leucine-rich repeat and C-terminal WD-40 domains. Pathogenic LRRK2 variants result in LRRK2 kinase activation with subsequent hyperphosphorylation of its endogenous substrates, a subgroup of RabGTPases including Rab10 (eLife 2016;5:e12813). Measuring LRRK2 dependent phosphorylation of Rab10 at Threonine 73 correlates with LRRK2 kinase activity (eLife 2016;5:e12813).

Here, we used a heterologous transient overexpression system of PD associated LRRK2 variants in HEK293 cells to measure their direct effect on phosphorylation levels of Rab10 at Threonine 73 (pRab10Thr73) as a readout for LRRK2 kinase pathway activation status in comparison to LRRK2 wildtype (Alexia’s paper). For each variant, at least 4 independent biological replicate experiments were performed for quantification. Additionally, the estimated conservation score for each LRRK2 variant is provided (PMID: 27166375).

Interpretation: LRRK2 variants with 1.5-fold activation of LRRK2 kinase pathway activity (>50% higher levels of phosphorylated pRab10Thr73 compared to LRRK2 wildtype) were considered “activating LRRK2 variants in a cellular transient overexpression assay ” with

  • pRab10Thr73 levels <1.5 and > 2-fold as “mildly activating”
  • <2 and >3 as “moderately activating” and
  • < 3 as “strongly activating”.

Variants with pRab10Thr73 levels < 0.5 > 1.5 compared to wildtype were considered “not activating in a cellular assay” and those with pRab10Thr73 levels >0.5 as “reduced activity in a cellular assay”.

LRRK2 dependent pRab10Thr73 phosphorylation can also be measured in human bio-samples such as neutrophils and monocytes isolated from fresh peripheral blood “in vivo” (PMID: 34125248, PMID: 29127255). For each variant, information on availability on patient data for LRRK2 dependent pRab10Thr73 phosphorylation is given and if applicable the interpretation thereof.

For example, the common G2019S variant results in “in vitro mild activation” (>1.75-fold above LRRK2 wild type) of LRRK2 kinase pathway activity, but its effect is “not statistically significant in an unpaired t-test”. Patient derived data exists, but LRRK2 dependent pRab10Thr73 phosphorylation is “not significantly elevated in vivo” when compared to controls in human derived peripheral blood neutrophils (PMID: 34125248). On the other hand, the R1441G variants activates LRRK2 kinase pathway activity “in vitro strongly”, 3.9-fold and patient derived data confirms “significantly elevated” LRRK2 kinase pathway activity in human peripheral blood neutrophils “in vivo” as well when compared to healthy controls (PMID: 34125248). Based on our current knowledge, we would expect that significant LRRK2 kinase pathway activation can be shown in any variant that activates LRRK2 kinase pathway activity in vitro strongly, above 3-fold compared to the LRRK2 wildtype protein.

Abbreviations/conventions used throughout MDSGene

AAO = age at onset
comp. het. = compound heterozyous
het = heterozygous
hom = homozyogus
N = number
n.a. = not applicable
O = other/mixed (ethnicity)
SD = standard deviation

Note that the full names of the official gene names can be found in the EntrezGene database. In addition, country names have been abbreviated according to the official 3-letter codes recommended by the International Organization for Standardization (ISO).


Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat Genet 2014;46:310–5

Kasten M, Hartmann C, Hampf J, Schaake S, Westenberger A, Vollstedt EJ, Balck A, Domingo A, Vulinovic F, Dulovic M, Zorn I, Madoev H, Zehnle H, Lembeck CM, Schawe L, Reginold J, Huang J, König IR, Bertram L, Marras C, Lohmann K, Lill CM, Klein C. Genotype-Phenotype Relations for the Parkinson's Disease Genes Parkin, PINK1, DJ1: MDSGene Systematic Review. Mov Disord 2018;33:730-41.

Lill CM, Mashychev A, Hartmann C, Lohmann K, Marras C,. Lang AE, Klein C, Bertram L . Launching the Movement Disorders Society Genetic Mutation Database (MDSGene). Mov Disord, 2016 May;31(5):607-9

Marras C, Lang A, van de Warrenburg BP, Sue CM, Tabrizi SJ, Bertram L, Mercimek-Mahmutoglu S, Ebrahimi-Fakhari D, Warner TT, Durr A, Assmann B, Lohmann K, Kostic V, Klein C. Recommendations of the International Parkinson and Movement Disorder Society Task Force on Nomenclature of Genetic Movement Disorders. Mov Disord 2016;31(4):436-57.

Rentzsch P, Witten D, Cooper GM, Shendure J, Kircher M. CADD: predicting the deleteriousness of variants throughout the human genome.Nucleic Acids Res. 2018 Oct 29.