Gene4PD: a comprehensive genetic database of Parkinson's disease

doi:10.21203/rs.3.rs-96276/v1

Download PDF

Research

Gene4PD: a comprehensive genetic database of Parkinson's disease

https://doi.org/10.21203/rs.3.rs-96276/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 26 Apr, 2021

Read the published version in Frontiers in Neuroscience →

Version 1

posted

You are reading this latest preprint version

Background

Parkinson's disease (PD) is a complex neurodegenerative disorder with a strong genetic component. A growing number of variants and genes have been reported to be associated with PD; however, there is no database that integrate different type of genetic data, and support analyzing of PD-associated genes (PAGs).

Methods

By systematic review and curation of multiple lines of public studies, we integrate multiple layers of genetic data (rare variants and copy-number variants identified from patients with PD, associated variants identified from genome-wide association studies, differential expression genes, and differential DNA methylation genes) and clinical data in PD. A weighted scoring system was employed to prioritize PAGs. Permutation test and protein-protein interaction network was used to evaluate the interconnectivity and functional correlation among the PAGs. The relationship between AAO and PAGs was further analyzed. A PHP-based web framework was used to construct a database.

Result

We integrated five layers of genetic data with different levels of evidences from more than 3,000 studies and prioritized 124 PAGs with strong or suggestive evidences. These PAGs were identified to be significantly interacted with each other and formed an interconnected functional network enriched in several functional pathways involved in PD, suggesting these genes may contribute to the pathogenesis of PD, which also highlighting the reliability of these genetic data for PD. Furthermore, we identified 10 genes were associated with a juvenile-onset (age ≤ 30 years), 11 genes were associated with an early-onset (age of 30–50 years), whereas another 10 genes were associated with a late-onset (age > 50 years). Notably, the AAOs of patients with loss of function variants in five genes (GCH1, PINK1, PRKN, FBXO7, ATP13A2) were significantly lower than that of patients with deleterious missense variants, while patients with VPS13C (P = 0.01) was opposite. Finally, we developed an online database named Gene4PD (http://genemed.tech/gene4pd) which integrated published genetic data in PD, the PAGs, and 63 popular genomic data sources, as well as an online pipeline for prioritize risk variants in PD.

Conclusion

Gene4PD provides researchers and clinicians comprehensive genetic knowledge and analytic platform for PD, and would also improve the understanding of pathogenesis in PD.

Availability and Implementation: Gene4PD can be freely accessed at http://genemed.tech/gene4pd.

Translational Medicine

Cellular & Molecular Neuroscience

Parkinson’s disease

PD-associated genes

Gene4PD database

age at onset

genetic variant

Parkinson’s disease (PD) is the second most common neurodegenerative disease(Dorsey et al., 2007; Alzheimer's, 2014). PD, particularly early-onset disease, shows high heritability(Blauwendraat et al., 2019a), suggesting that genetic factors play important roles in its pathogenesis. In the last two decades with the development of high throughput technologies including next-generation sequencing(Jansen et al., 2017; Sandor et al., 2017; Siitonen et al., 2017), numerous studies have focused on the genetic architectures to understand the pathogenesis of PD(Kalia and Lang, 2015; Labbe et al., 2016; Blauwendraat et al., 2019b). Several rare variants and candidate genes have been identified among familial and sporadic cases of PD(Funayama et al., 2015; Kalia and Lang, 2015; Deng et al., 2016; Farlow et al., 2016; Lesage et al., 2016; Mok et al., 2016; Guo et al., 2018; Quadri et al., 2018; Blauwendraat et al., 2019b). Previously, we have replicated PRKN(Guo et al., 2008), PINK1(Tang et al., 2006; Guo et al., 2008), PARK7(Tang et al., 2006; Guo et al., 2008), ATP13A2(Guo et al., 2008), PLA2G6(Shi et al., 2011), CHCHD2(Liu et al., 2015), RAB39B(Kang et al., 2016), TMEM230(Yan et al., 2017), GCH1(Xu et al., 2017), and other genes(Tian et al., 2012; Yang et al., 2019) in patients with PD in China. Copy number variations (CNVs), particularly in PRKN and SNCA, have also been highlighted to play a vital role in the development of heritable or sporadic PD(Konno et al., 2016; Mok et al., 2016; Taghavi et al., 2018) and several genes, such as ATP13A2(Ramirez et al., 2006), PLA2G6(Paisan-Ruiz et al., 2009), FBXO7(Di Fonzo et al., 2009), PRKN(Lucking et al., 2000; Djarmati et al., 2004), PINK1(Bonifati et al., 2005; Kumazawa et al., 2008), and PARK7(Djarmati et al., 2004) were reported to be strongly associated with an early-onset age of PD(Searles Nielsen et al., 2013).

In addition, genome-wide association studies (GWAS) have provided insight into the genetic basis of PD by identifying and replicating risk loci(Chung et al., 2012; Nalls et al., 2014; Chang et al., 2017; Foo et al., 2017; Jansen et al., 2017; Blauwendraat et al., 2019b). Some differentially DNA methylated genes(Wullner et al., 2016) and differentially expressed genes(Infante et al., 2015) have also been reported to be associated with PD pathogenesis. Most currently known PD-causing genes are involved in several pathogenesis pathways including mitochondrial dysfunction, neuron apoptosis, and the autophagy lysosome ubiquitin-proteasome system(Blauwendraat et al., 2019b). Despite these great advances in the genetics and genomics of PD, the related results are scattered among thousands of studies, making it difficult and laborious for researchers to use these available data.

PD is an age-dependent neurodegenerative condition, and the age at onset (AAO) is a robust phenotypic measure compared to motor and nonmotor phenotypes (Wickremaratchi et al., 2011; Kilarski et al., 2012; Nalls et al., 2015). Many previous studies have reported the frequency of gene mutations in PD patients with different AAOs, and most of them focused on analyzing the association of AAO and several common PD-causing genes(Searles Nielsen et al., 2013; Kasten et al., 2018; Trinh et al., 2018; Tan et al., 2019). Up to now, several genes has been reported to be strongly associated with an early-onset age of PD based on cohorts within specialist clinics or within the general population(Searles Nielsen et al., 2013; Kasten et al., 2018; Trinh et al., 2018; Lin et al., 2019; Tan et al., 2019; Trinh et al., 2019). However, a comprehensive analysis the association between AAO and all PD-associated genes (PAGs) is still lacking. Furthermore, where AAO is associated other genetic components, such as the expression pattern, the functional effects of genetic variants need to be further investigated.

To overcome this limitation and facilitate studies of the genetic basis of PD, we integrated multiple layers of genetic and genomic data of PD and prioritized several PAGs, we developed a convenient integrative genomic database named as Gene4PD, which can facilitate queries and analysis of genetic and genomic data with an intuitive graphical user interface that will benefit physicians and researchers.

Data collection

We collected five types of genomic data related to PD from the PubMed database up to August 20, 2020 with following search items: (1) rare variants identified from PD patients with “Parkinson*[Title/Abstract] AND (mutation [Title/Abstract] or variant [Title/Abstract]) AND gene [Title/Abstract]”; (2) CNVs in PD with “Parkinson*[Title/Abstract] AND (copy number variation [Title/Abstract] or CNV [Title/Abstract])”; (3) associated SNPs identified from GWAS with “Parkinson*[Title/Abstract] AND GWAS[Title/Abstract]”; (4) differential expressed genes (DEGs) with “Parkinson*[Title/Abstract] AND transcriptome [Title/Abstract]”; (5) differentially expressed DNA methylation genes (DMGs) with “Parkinson*[Title/Abstract] AND DNA methylation [Title/Abstract]”. We then manually screened out articles related to PD and extracted available genomic and clinical information based on the uniform criteria outlined in Table S1.

Data annotation and PAG prioritization

To investigate the functional genetic variants identified in patients with PD, ANNOVAR(Wang et al., 2010) was used for comprehensive annotation based on definitions of transcripts from the RefSeq database. Based on the functional effects and minor allele frequency (MAF) in the GnomAD database, the genetic variants were classified as follows: (1) rare loss of function (LoF, including frameshift indels, splicing, stop-gain, and stop-loss) variants with MAF less than 0.01; (2) rare deleterious missense (Dmis) variants with MAF less than 0.01; (3) rare tolerate missense (Tmis) variants with MAF less than 0.01; and (4) all remaining genetic variants. We used our previously developed program ReVe(Ioannidis et al., 2016) to predict deleterious missense variants with scores higher than 0.7, as previous studies.

To prioritize PAGs, we developed a scoring system (Table S2) combining the different types of genetic evidences as following. The rare variants assigned an evidence scores of 1–5 according to their functional effects and MAF. The CNVs were directly assigned an evidence score of 5 because they disrupted gene function. The PD-associated SNPs, DEGs, and DMGs were assigned an evidence scores of 1–3 based on their p-values. A combined evidence score for each gene was calculated by summing up the evidence scores of all integrated studies. All genes integrated in this study were classified into five grades: high confidence (score ≥ 20), strong associated (score of 10–20), suggestive associated (score of 5–10), minimal evidence (score of 3–5), and uncertain evidence (score < 3) (Fig. 1).

The verification of functional correlation and PAG network

We performed a permutation test to evaluate the interconnectivity and functional correlation among the 124 PAGs (score ≥ 5) as described in our previous study(Li et al., 2017). To ensure a reasonable level of analysis, we constructed a protein-protein interaction (PPI) network using the STRING v 11.0 database (https://string-db.org/)(Szklarczyk et al., 2019) with a confidence score greater than 0.4. Specifically, we randomly simulated 1,000,000 permutation tests to evaluate the interconnectivity among high-confidence/strong genes, and among suggestive genes, as well as to determine the connectivity between these two classes of PAGs. Then, the 124 PAGs were selected to construct an interconnected PPI network based on the STRING online analysis platform. Additionally, the functional networks were clustered by multiple biological processes of Gene Ontology (GO) (http://www.geneontology.org/).

Association between age at onset and PAGs

The AAO data and genetic data of each gene were download from Gene4PD. We then analyzed the relationship between the AAO and PAGs from three perspectives. First, we characterized the association of AAO and spatio-temporal expression patterns by comparing the median AAO of genes in a different module, which was achieved by performing a WGCNA, as described above. Second, we aggregated and compared genes with more than five AAO terms to assess the association between AAO and PAGs. All genes were classified into one of three sections: juvenile-onset (≤ 30 years), early-onset (30–50 years), or late-onset (> 50 years), according to the median of the AAO. Third, we compared the AAO with different functional types of variants (loss-of-function [LoF] and deleterious missense variants [Dmis]) in the same gene to investigate the contribution of different types of functional variants (LoF and Dmis) to the AAO of PD.

Database construction and interface

By integrating all the collected genomic information as well as the annotation for each variant/gene from 63 data sources (Table S3), we developed Gene4PD database (http://www.genemed.tech/gene4pd/) by combining Vue with a PHP-based web framework lavarel to construct as a user-friendly web interface. Moreover, the front and back separation model was used for website development. The front end is based on vue and uses the UI Toolkit element, which supports all modern browsers across platforms, including Microsoft Edge, Safari, Firefox, and Google Chrome, and the back end is based on laravel, a PHP web framework. Gene4PD is compatible with all major browser environments and different operating systems such as Windows, Linux, and Mac. The data are stored in the MySQL database.

Genomic data integration and PAG prioritization

Through systematic review and curation of multiple lines of public studies, we reviewed more than 3,000 publications, 487 of which met the quality control and collection criteria. Genetic information such as gene symbols, chromosomes, locations, reference base, altered base, and hereditary modes were collected, and other basic information and clinical data, including sample ID, PubMed ID, methods of detecting variants, country, race, gender, AAO, functional study, subtype of disease, detail description of clinical phenotypes, and sporadic/familiar types, were also integrated. We catalogued five types of genetic data related to PD: (1) 2,252 rare variants, including 954 nonredundant rare variants, in 226 genes from 327 available publications; (2) 139 CNVs in 34 genes from 94 publications; (3) 1,237 associated SNPs in 640 genes from 42 studies; (4) 2,926 DEGs from 8 publications; (5) 657 DMGs from 7 publications (Table S1) (Fig. 1). Specifically, for the collected genetic variants, we identified 334 rare LoF, 1,328 rare Dmis variants, 485 rare Tmis variants, and 105 other remaining variants based on standard annotations.

We then developed a weighted scoring system to prioritize PAGs by combining all above genetic evidence. As a result, we prioritized 25 high confidence genes (score ≥ 20), 38 strong associated genes (score of 10–20), 61 suggestive associated genes (score of 5–10), 88 minimal evidence genes (score of 3–5), and 3,791 uncertain evidence genes (score < 3) (Table 1). We found that 19 of 21 known PD-causing genes (Table S4) were prioritized as high confidence genes. Six other high confidence genes, including GCH1, RAB39B, CHCHD2, MAPT, TH, and ASNA1, were also regarded as potential PD-causing genes. The remaining two known PD-causing genes were classified into strong associated genes (HTRA2) or suggestive associated genes (UCHL1), as they showed lower replication.

PAGs were functionally correlated

We performed a permutation test to assess the functional correlations of the 124 associated genes (score ≥ 5). Most of them were disease-causing genes or risk-genes from GWAS. As a result, we observed 48 of 63 high-confidence or strong PAGs (P < 1.0 × 10^− 6, Fig. S1a) that interacted with each other and had 203 interconnections (P < 1.0 × 10^− 6, Fig. S1b), which was significantly higher than the random expectation. Similarly, the suggestive associated genes also significantly interacted with each other (Fig. S1c–d), suggesting they were functionally correlated. Furthermore, we observed 21 suggestive associated genes (P = 0.098, Fig. S1e) which interacted with high-confidence or strong PAGs with 74 connections (P = 2.5 × 10^− 5, Fig. S1f), suggesting that the two classes of associated genes were also functionally correlated. These results demonstrate that the 124 genes were functionally associated with PD, although these results require further experimental validation.

We then developed an interacted functional network contained 88 of the 124 PAGs which interacted with each other at protein-level with 336 connections (Fig. 2a). The functional network contained 25 high-confidence genes, 27 strong associated genes, and 36 suggestive associated genes. All 21 known PD-causing genes were included in this functional network. Additionally, other genes in this PPI network may be associated with PD. GO enrichment analysis of the 88 genes revealed server GO associated with PD (Table S5, Fig. 2b). Most of these GO terms, such as regulation of neuron death (GO:1901214, FDR = 7.81 × 10^− 10), regulation of autophagy (GO:0010506, FDR = 2.30 × 10^− 8), behavior (GO:0007610, FDR = 4.46 × 10^− 8), regulation of cellular catabolic process (GO:0031329, FDR = 8.84 × 10^− 8), (Table S5, Fig. 2b), were regarded as critical functional signaling pathways associated with PD. The functional network suggested that the prioritized PAGs shared a common signaling mechanism and were functionally correlated.

AAO is associated with multiple genetic components

The integrated genetic data and AAO data from the Gene4PD database provide an unprecedented opportunity to comprehensively identify the vital association between the AAO and PAGs on a large scale. Therefore, we analyzed 31 PAGs with more than five AAO items in each gene (Fig. 3a). After sorting by the median AAO of each gene, we found that 10 genes (SLC6A3, TH, L2HGDH, SLC30A10, DNAJC6, GCH1, FBXO7, ATP13A2, SYNJ1, PLA2G6) were associated with a juvenile-onset (age ≤ 30 years), 11 genes (PRKN, PARK7, PINK1, HTRA2, VPS13C, GBA, ATP10B, POLG, CHCHD2, SNCA, RAB39B) were associated with an early-onset (age of 30–50 years), whereas another 10 genes (VPS35, DCTN1, LRRK2, MAPT, TMEM230, GIGYF2, DNAJC13, TARDBP, LRP10, CSMD1) were associated with a late-onset (age > 50 years). Although different PAGs had specific AAO characteristics, we noted that there were large differences in the AAO for some genes, such as for PRKN, PINK1, SNCA, and LRRK2, suggesting that there are other factors that contribute to the AAO.

To further investigate different types of functional variants (LoF and Dmis) contributing to the AAO of PD, we further performed pairwise comparison to analyze the differences in the AAO between patients with LoF and patients with Dmis for each gene. As a result, 10 genes with more than three AAO items both in LoF and Dmis were selected to perform a comparison analysis. We found that the AAO of patients with LoF in five genes, including GCH1 (P = 2.63 × 10^− 5), PINK1 (P = 2.31 × 10^− 3), PRKN (P = 4.08 × 10^− 3), FBXO7 (P = 0.02), and ATP13A2 (P = 0.03), were significantly lower than patients with Dmis, while the AAO of patients with LoF in VPS13C (P = 0.01) was higher than patients with Dmis, whereas four other genes including PARK7 (P = 0.06), SNCA (P = 0.12), RAB39B (P = 0.96), and TMEM230 (P = 0.27) did not present significant differences (Fig. 3b). More PD samples with detailed genotypes and phenotypes may be identified in further studies investigating the relationship between AAO and PAGs.

Gene4PD: an integrative genetic database and analytic platform for Parkinson’s disease

To facilitate researchers to query and analyze genetic and genomic data related to PD, we constructed an online comprehensive database and analysis platform named as Gene4PD (http://www.genemed.tech/gene4pd/) which integrated rare variants, CNVs, associated SNPs, DEGs, and DMGs related PD and more than 63 popular genomic and genetic data sources (Fig. 1, Table S3). Users can query the detailed genetic and genomic information of the variants and genes via quick or advanced search interfaces in Gene4PD (Fig. 4a). The quick search supports several common search terms (such as gene symbol, genomic regions, cytoband) and returns a page containing six visual tables. The advanced search in Gene4PD supports additional user demands by allowing the user to upload a file or paste a list containing search terms. The first table summarizes the evidences scores from the variant types of the collected genetic data, and the other five tables show the more detailed information (Fig. 4b). In addition, to comprehensively evaluate the pathogenicity of genetic variants, three meaningful panels were integrated in Gene4PD: (1) predicted damaging scores and functional consequences of missense variants from 24 in silico algorithms (Table S3); (2) allele frequencies of different populations based on seven databases, including gnomAD, ExAC, 1000 Genomes Project, ESP6500(Fu et al., 2013), Kaviar, and Haplotype Reference Consortium; (3) disease-related information from 11 related databases, including InterVar, COSMIC70, ICGC, dbSNP, ClinVar, Gene4Denovo, InterPro, OMIM, MGI, and HPO. Notably, all query results can be copied to the clipboard or exported as Excel spreadsheets.

Gene4PD includes six sections to obtain gene-level knowledge of a given gene in a one-stop interface (Fig. S2). (1) The Basic information sections gives out the primary information of genes sourced from NCBI Gene, Entrez gene, Ensembl, OMIM, GeneCards, HGNC, the intolerance score from residual variation intolerance score (RVIS), LoFtool, heptanucleotide context intolerance score, GDI, Episcore, and pLI score; (2) The Gene function section includes five sub-sections: molecular function extracted from UniportKB, GO terms, Domian from InterPro, PPI from InBio Map, biological pathway from Biosystems; (3) The Phenotype and Disease section reports phenotype and disease-related variants or genes from the OMIM, ClinVar, MGI, HPO and Gene4Denovo are shown in the Phenotype and disease section; (4) The Gene expression section shows spatiotemporal-expression levels of genes in the brain retrieved from BrianSpan, expression profiles in 31 primary tissues and in 54 secondary tissues extracted from GTEx, cell diversity and expression in the human cortex based on single-nucleus RNA-seq data from Allen Brain Atlas, and the subcellular location retrieved from The Human Protein Atlas; (5) The Variants in different populations section provides the number of variants with functional effects in different populations from genomAD; (6) The Drug-gene interaction section provides the drug-gene interactions and gene druggability data sourced from DGIdb v3.0.2.

One of the main advantages of Gene4PD is that it provides an interface for analyzing genetic data according to the specific needs of the user (Fig. S3). The analysis process includes four simple steps of filling in an email address, choosing the Trio or Non-trio option of the genetic data, uploading genetic data files (VCF4 format), and inputting basic information of samples. Gene4PD will then analyze the rare damaging variants using default parameters. Importantly, Gene4PD also supports a flexible control panel so that users can conveniently adjust the values of quality control, data sources of annotation, and parameters for identifying rare damaging variants. The annotation section contains four sub-panels: Basic information annotation, Pathogenicity prediction of missense variants, Allele frequency in variant population, and Clinical-related database. When the analysis is complete, Gene4PD will send a link to the user by email for downloading the results.

With the development of NGS, the number of multi-omics studies has greatly increased, such as genomics, genetics, epigenetics, and transcriptomics, for revealing the pathogenesis of PD, resulting in the identification of numerous associated genes(Jansen et al., 2017; Sandor et al., 2017; Siitonen et al., 2017). The ability to comprehensively investigate these associated genes from different genetic perspectives such as expression patterns, functional interconnections, and relationships with AAO would be helpful for studying the pathogenesis of PD, and improving clinical diagnosis, treatment, and drug development efforts(Blauwendraat et al., 2019b). However, most of these results are widely scattered among published articles, making it challenging to integrate these omics data. To facilitate this task, in this study, we manually collected rare variants, CNVs, associated SNPs, DEGs, and DMGs related to PD from the related literature, and prioritized 124 PAGs, including 25 high-confidence, 38 strong associated and 61 suggestive associated genes. Interestingly, 19 of the 21 known PD-causing genes belonged to the high-confidence genes, highlighting that other high-confidence PAGs may also be associated with PD. Therefore, more genetic and experimental studies are needed to validate the genetic mechanisms of PAGs incorporating all of these associations.

We further performed several bioinformatics analyses to validate the functional association of the 124 PAGs based on PPI data. First, we observed significant associations among the high-confidence/strong PAGs, among the suggestive PAGs, and between the two classes of experimental evidences. Second, we developed an interconnected functional network containing 88 PAGs and enriched in several known PD-associated functional pathways. In the functional network, we highlighted several PAGs that interacted with known PD-causing genes, such as CHCHD2, RAB39B, and RIC3. CHCHD2(Funayama et al., 2015) has been widely reported to be associated with PD, presented functional interaction with many PD-causing genes, such as SNCA, PINK1, LRRK2, PARK7, VPS35 in the network. RAB39B with the known PD-causing genes SNCA and LRRK2 is consistent with its α-synuclein pathology(Wilson et al., 2014). These results suggest that these three genes are involved in closely related biological functions similar to the disease-causing genes and increasing the risk PD.

To enable PD researchers to make full use of the reported genetic data, more than six online PD-related database such as the variation databases MDSGene(Lill et al., 2016), PDGene database(Nalls et al., 2014), ParkDB(Taccioli et al., 2011), Parkinson Disease Mutation Database (PDmutDB)(Horaitis et al., 2007), The Mutation Database for PD (MDPD)(Tang et al., 2009), (Parkinson’s disease map) PDMap(Fujita et al., 2014) and PDbase(Yang et al., 2009) have been explored. However, we found that MDSGene collected mutations of 12 common PD associated genes, and the other three databases (PDmutDB, MDPD, PDbase) were not in functional order for use. In addition, the PDGene database focused on GWAS data collection and analysis and did not provide summary and statistic reports, whereas ParkDB is specifically dedicated to gene expression data and PDmap integrates pathways implicated in PD pathogenesis. Therefore, to establish an integrative genetic database and analytic platform for facilitating PD research, we integrated all collected data to construct Gene4PD as a comprehensive data analysis platform. Compared to existing databases, Gene4PD not only integrates more recorded variants, CNVs, SNPs from GWAS, DNA methylation genes, and DEGs, but also supports the systematic analysis of genetic variants. To overcome the multiple challenges in evaluating the pathogenicity of a variant or to determine whether a gene is pathogenic, based on experience with the construction of our previous online databases VarCards(Li et al., 2018) and Gene4Denovo(Guihu et al., 2019), we integrated more than 63 genomic data resources in Gene4PD, covering comprehensive variant-level and gene-level annotation data. In addition, we will update Gene4PD by collecting published gene-related data every six months.

PD has been widely recognized as an age-dependent neurodegenerative condition. The variation in phenotypes and genotypes can be related to the AAO of PD(Wickremaratchi et al., 2011; Kilarski et al., 2012; Nalls et al., 2015). Previous studies suggested that some genes were strongly associated with the AAO, such as ATP13A2(Ramirez et al., 2006), PLA2G6(Paisan-Ruiz et al., 2009), FBXO7(Di Fonzo et al., 2009), PRKN(Lucking et al., 2000; Djarmati et al., 2004), PINK1(Bonifati et al., 2005; Kumazawa et al., 2008), and PARK7(Djarmati et al., 2004). Based on our largescale integrated dataset, we not only confirmed the association between these six commonly reported genes and AAO, but also verified another 20 genes associated with AAO. Furthermore, our results suggest that different types of functional variants (LoF and Dmis) contribute differently to the AAO of PD.

There were some limitations to this study. First, by satisfying the quality control, we collected as much related literature as possible, but there may have been some omissions. Second, the scoring system used in this study may not be powerful enough, and thus we encourage users to analyze genetic data with the re-score function in Gene4PD or to download data for re-analysis. Third, the integrated approach in this study may be biased, and the prioritized genes need to be validated in population and verified the pathogenic mechanism with cell or animal experiments. Last, clinical information besides the AAO must be considered to determine the factors contributing to the genotype-phenotype correlations of PD. However, most articles do not provide detailed phenotypic information, which poses a challenge for a meta-analysis of the relationship between clinical phenotypes and genes. In this regard, we encourage researchers to provide details on clinical phenotypes in the context of genetic studies.

In conclusion, we catalogued different types of genetic data from many publications related to PD and prioritized 124 functionally related PAGs with different lines of evidence, which were used to construct the database and analysis tool Gene4PD, suggesting that integrating multiple genetic data is useful for prioritizing novel associated genes. In addition, we characterized the genetic landscape of the prioritized PAGs, providing insight into the pathology of PD. Gene4PD is expected to accelerate genetic consulting, subtype definition, and drug development in PD. Moreover, as an integrative genetic database and analytic platform, Gene4PD can be used by geneticists and clinicians to obtain more in-depth genetic knowledge of PD.

PD: Parkinson’s disease; AAO = age at onset; DEGs = differential expressed genes; DMGs = differentially expressed DNA methylation genes; GO = Gene Ontology; GWAS = genome-wide association studies; MAF = minor allele frequency; PAGs = Parkinson’s disease-associated genes;

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

The datasets analyzed during the current study are available in the Gene4PD repository (http://genemed.tech/gene4pd/).

Competing interests

The authors declare no competing interests.

Funding

This work was financially supported by National Natural Science Foundation of China [grant number 81430023 to BST, 82001362 to BL]; the National Key Plan for Scientific Research and Development of China grants [grant number 2016YFC1306000 to BST]; Young Elite Scientist Sponsorship Program by CAST [grant number 2018QNRC001 to JCL]; and Innovation-Driven Project of Central South University [grant number 20180033040004 to JCL];.

Authors' contributions

Study design: BL, GHZ, BST, JCL;

Literature search and data collection: BL, GHZ, QZ, YLX, ZW, ZHF, BLu, LXQ, YWZ, RZ, LJ, HXP, YH, XMW, TFL, YZ, YJW, QC, and JCL;

Figures: BL, GHZ, QZ, YLX, BST, JCL

Data analysis: BL, GHZ, ZHF, BLu, XMW, TFL, YZ, ZHL, JFG and JCL;

Data interpretation: BL, GHZ, ZW, YJW, QC, BST, and JCL;

Writing of the manuscript draft and all revision stages: BL, GHZ, QZ, YLX, ZW, ZHF, BLu, LXQ, YWZ, RZ, LJ, HXP, YH, XMW, TFL, YZ, YJW, QC, ZHL, JFG, BST and JCL.

Acknowledgements

This project was funded by the Central South University. The funder provided support in the form of salaries of all authors.

Alzheimer's A. 2014 Alzheimer's disease facts and figures. Alzheimers Dement 2014; 10(2): e47-92.

Blauwendraat C, Heilbron K, Vallerga CL, Bandres-Ciga S, von Coelln R, Pihlstrom L, et al. Parkinson's disease age at onset genome-wide association study: Defining heritability, genetic loci, and alpha-synuclein mechanisms. Mov Disord 2019a; 34(6): 866-75.

Blauwendraat C, Nalls MA, Singleton AB. The genetic architecture of Parkinson's disease. Lancet Neurol 2019b.

Bonifati V, Rohe CF, Breedveld GJ, Fabrizio E, De Mari M, Tassorelli C, et al. Early-onset parkinsonism associated with PINK1 mutations: frequency, genotypes, and phenotypes. Neurology 2005; 65(1): 87-95.

Chang D, Nalls MA, Hallgrimsdottir IB, Hunkapiller J, van der Brug M, Cai F, et al. A meta-analysis of genome-wide association studies identifies 17 new Parkinson's disease risk loci. Nat Genet 2017; 49(10): 1511-6.

Chung SJ, Armasu SM, Biernacka JM, Anderson KJ, Lesnick TG, Rider DN, et al. Genomic determinants of motor and cognitive outcomes in Parkinson's disease. Parkinsonism Relat Disord 2012; 18(7): 881-6.

Deng HX, Shi Y, Yang Y, Ahmeti KB, Miller N, Huang C, et al. Identification of TMEM230 mutations in familial Parkinson's disease. Nat Genet 2016; 48(7): 733-9.

Di Fonzo A, Dekker MC, Montagna P, Baruzzi A, Yonova EH, Correia Guedes L, et al. FBXO7 mutations cause autosomal recessive, early-onset parkinsonian-pyramidal syndrome. Neurology 2009; 72(3): 240-5.

Djarmati A, Hedrich K, Svetel M, Schafer N, Juric V, Vukosavic S, et al. Detection of Parkin (PARK2) and DJ1 (PARK7) mutations in early-onset Parkinson disease: Parkin mutation frequency depends on ethnic origin of patients. Hum Mutat 2004; 23(5): 525.

Dorsey ER, Constantinescu R, Thompson JP, Biglan KM, Holloway RG, Kieburtz K, et al. Projected number of people with Parkinson disease in the most populous nations, 2005 through 2030. Neurology 2007; 68(5): 384-6.

Farlow JL, Robak LA, Hetrick K, Bowling K, Boerwinkle E, Coban-Akdemir ZH, et al. Whole-Exome Sequencing in Familial Parkinson Disease. JAMA Neurol 2016; 73(1): 68-75.

Foo JN, Tan LC, Irwan ID, Au WL, Low HQ, Prakash KM, et al. Genome-wide association study of Parkinson's disease in East Asians. Hum Mol Genet 2017; 26(1): 226-32.

Fu W, O'Connor TD, Jun G, Kang HM, Abecasis G, Leal SM, et al. Analysis of 6,515 exomes reveals the recent origin of most human protein-coding variants. Nature 2013; 493(7431): 216-20.

Fujita KA, Ostaszewski M, Matsuoka Y, Ghosh S, Glaab E, Trefois C, et al. Integrating pathways of Parkinson's disease in a molecular interaction map. Mol Neurobiol 2014; 49(1): 88-102.

Funayama M, Ohe K, Amo T, Furuya N, Yamaguchi J, Saiki S, et al. CHCHD2 mutations in autosomal dominant late-onset Parkinson's disease: a genome-wide linkage and sequencing study. Lancet Neurol 2015; 14(3): 274-82.

Guihu Z, Kuokuo L, Bin L, Zheng W, Zhenghuan F, Xiaomeng W, et al. Gene4Denovo: an integrated database and analytic platform for de novo mutations in humans. Nucleic Acids Research 2019.

Guo JF, Xiao B, Liao B, Zhang XW, Nie LL, Zhang YH, et al. Mutation analysis of Parkin, PINK1, DJ-1 and ATP13A2 genes in Chinese patients with autosomal recessive early-onset Parkinsonism. Mov Disord 2008; 23(14): 2074-9.

Guo JF, Zhang L, Li K, Mei JP, Xue J, Chen J, et al. Coding mutations in NUS1 contribute to Parkinson's disease. Proc Natl Acad Sci U S A 2018; 115(45): 11567-72.

Horaitis O, Talbot CC, Jr., Phommarinh M, Phillips KM, Cotton RG. A database of locus-specific databases. Nat Genet 2007; 39(4): 425.

Infante J, Prieto C, Sierra M, Sanchez-Juan P, Gonzalez-Aramburu I, Sanchez-Quintana C, et al. Identification of candidate genes for Parkinson's disease through blood transcriptome analysis in LRRK2-G2019S carriers, idiopathic cases, and controls. Neurobiol Aging 2015; 36(2): 1105-9.

Ioannidis NM, Rothstein JH, Pejaver V, Middha S, McDonnell SK, Baheti S, et al. REVEL: An Ensemble Method for Predicting the Pathogenicity of Rare Missense Variants. Am J Hum Genet 2016; 99(4): 877-85.

Jansen IE, Ye H, Heetveld S, Lechler MC, Michels H, Seinstra RI, et al. Discovery and functional prioritization of Parkinson's disease candidate genes from large-scale whole exome sequencing. Genome Biol 2017; 18(1): 22.

Kalia LV, Lang AE. Parkinson's disease. Lancet 2015; 386(9996): 896-912.

Kang JF, Luo Y, Tang BS, Wan CM, Yang Y, Li K, et al. RAB39B gene mutations are not linked to familial Parkinson's disease in China. Sci Rep 2016; 6: 34502.

Kasten M, Hartmann C, Hampf J, Schaake S, Westenberger A, Vollstedt EJ, et al. Genotype-Phenotype Relations for the Parkinson's Disease Genes Parkin, PINK1, DJ1: MDSGene Systematic Review. Mov Disord 2018; 33(5): 730-41.

Kilarski LL, Pearson JP, Newsway V, Majounie E, Knipe MD, Misbahuddin A, et al. Systematic review and UK-based study of PARK2 (parkin), PINK1, PARK7 (DJ-1) and LRRK2 in early-onset Parkinson's disease. Mov Disord 2012; 27(12): 1522-9.

Konno T, Ross OA, Puschmann A, Dickson DW, Wszolek ZK. Autosomal dominant Parkinson's disease caused by SNCA duplications. Parkinsonism Relat Disord 2016; 22 Suppl 1: S1-6.

Kumazawa R, Tomiyama H, Li Y, Imamichi Y, Funayama M, Yoshino H, et al. Mutation analysis of the PINK1 gene in 391 patients with Parkinson disease. Arch Neurol 2008; 65(6): 802-8.

Labbe C, Lorenzo-Betancor O, Ross OA. Epigenetic regulation in Parkinson's disease. Acta Neuropathol 2016; 132(4): 515-30.

Lesage S, Drouet V, Majounie E, Deramecourt V, Jacoupy M, Nicolas A, et al. Loss of VPS13C Function in Autosomal-Recessive Parkinsonism Causes Mitochondrial Dysfunction and Increases PINK1/Parkin-Dependent Mitophagy. Am J Hum Genet 2016; 98(3): 500-13.

Li J, Shi L, Zhang K, Zhang Y, Hu S, Zhao T, et al. VarCards: an integrated genetic and clinical database for coding variants in the human genome. Nucleic Acids Res 2018; 46(D1): D1039-D48.

Li J, Wang L, Guo H, Shi L, Zhang K, Tang M, et al. Targeted sequencing and functional analysis reveal brain-size-related genes and their networks in autism spectrum disorders. Mol Psychiatry 2017; 22(9): 1282-90.

Lill CM, Mashychev A, Hartmann C, Lohmann K, Marras C, Lang AE, et al. Launching the movement disorders society genetic mutation database (MDSGene). Mov Disord 2016; 31(5): 607-9.

Lin CH, Chen PL, Tai CH, Lin HI, Chen CS, Chen ML, et al. A clinical and genetic study of early-onset and familial parkinsonism in taiwan: An integrated approach combining gene dosage analysis and next-generation sequencing. Mov Disord 2019; 34(4): 506-15.

Liu Z, Guo J, Li K, Qin L, Kang J, Shu L, et al. Mutation analysis of CHCHD2 gene in Chinese familial Parkinson's disease. Neurobiol Aging 2015; 36(11): 3117 e7- e8.

Lucking CB, Durr A, Bonifati V, Vaughan J, De Michele G, Gasser T, et al. Association between early-onset Parkinson's disease and mutations in the parkin gene. N Engl J Med 2000; 342(21): 1560-7.

Mok KY, Sheerin U, Simon-Sanchez J, Salaka A, Chester L, Escott-Price V, et al. Deletions at 22q11.2 in idiopathic Parkinson's disease: a combined analysis of genome-wide association data. Lancet Neurol 2016; 15(6): 585-96.

Nalls MA, Escott-Price V, Williams NM, Lubbe S, Keller MF, Morris HR, et al. Genetic risk and age in Parkinson's disease: Continuum not stratum. Mov Disord 2015; 30(6): 850-4.

Nalls MA, Pankratz N, Lill CM, Do CB, Hernandez DG, Saad M, et al. Large-scale meta-analysis of genome-wide association data identifies six new risk loci for Parkinson's disease. Nat Genet 2014; 46(9): 989-93.

Paisan-Ruiz C, Bhatia KP, Li A, Hernandez D, Davis M, Wood NW, et al. Characterization of PLA2G6 as a locus for dystonia-parkinsonism. Ann Neurol 2009; 65(1): 19-23.

Quadri M, Mandemakers W, Grochowska MM, Masius R, Geut H, Fabrizio E, et al. LRP10 genetic variants in familial Parkinson's disease and dementia with Lewy bodies: a genome-wide linkage and sequencing study. Lancet Neurol 2018; 17(7): 597-608.

Ramirez A, Heimbach A, Grundemann J, Stiller B, Hampshire D, Cid LP, et al. Hereditary parkinsonism with dementia is caused by mutations in ATP13A2, encoding a lysosomal type 5 P-type ATPase. Nat Genet 2006; 38(10): 1184-91.

Sandor C, Honti F, Haerty W, Szewczyk-Krolikowski K, Tomlinson P, Evetts S, et al. Whole-exome sequencing of 228 patients with sporadic Parkinson's disease. Sci Rep 2017; 7: 41188.

Searles Nielsen S, Bammler TK, Gallagher LG, Farin FM, Longstreth W, Jr., Franklin GM, et al. Genotype and age at Parkinson disease diagnosis. Int J Mol Epidemiol Genet 2013; 4(1): 61-9.

Shi CH, Tang BS, Wang L, Lv ZY, Wang J, Luo LZ, et al. PLA2G6 gene mutation in autosomal recessive early-onset parkinsonism in a Chinese cohort. Neurology 2011; 77(1): 75-81.

Siitonen A, Nalls MA, Hernandez D, Gibbs JR, Ding J, Ylikotila P, et al. Genetics of early-onset Parkinson's disease in Finland: exome sequencing and genome-wide association study. Neurobiol Aging 2017; 53: 195 e7- e10.

Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 2019; 47(D1): D607-D13.

Taccioli C, Tegner J, Maselli V, Gomez-Cabrero D, Altobelli G, Emmett W, et al. ParkDB: a Parkinson's disease gene expression database. Database (Oxford) 2011; 2011: bar007.

Taghavi S, Chaouni R, Tafakhori A, Azcona LJ, Firouzabadi SG, Omrani MD, et al. A Clinical and Molecular Genetic Study of 50 Families with Autosomal Recessive Parkinsonism Revealed Known and Novel Gene Mutations. Mol Neurobiol 2018; 55(4): 3477-89.

Tan MMX, Malek N, Lawton MA, Hubbard L, Pittman AM, Joseph T, et al. Genetic analysis of Mendelian mutations in a large UK population-based Parkinson's disease study. Brain 2019; 142(9): 2828-44.

Tang B, Xiong H, Sun P, Zhang Y, Wang D, Hu Z, et al. Association of PINK1 and DJ-1 confers digenic inheritance of early-onset Parkinson's disease. Hum Mol Genet 2006; 15(11): 1816-25.

Tang S, Zhang Z, Kavitha G, Tan EK, Ng SK. MDPD: an integrated genetic information resource for Parkinson's disease. Nucleic Acids Res 2009; 37(Database issue): D858-62.

Tian JY, Guo JF, Wang L, Sun QY, Yao LY, Luo LZ, et al. Mutation analysis of LRRK2, SCNA, UCHL1, HtrA2 and GIGYF2 genes in Chinese patients with autosomal dorminant Parkinson's disease. Neurosci Lett 2012; 516(2): 207-11.

Trinh J, Lohmann K, Baumann H, Balck A, Borsche M, Bruggemann N, et al. Utility and implications of exome sequencing in early-onset Parkinson's disease. Mov Disord 2019; 34(1): 133-7.

Trinh J, Zeldenrust FMJ, Huang J, Kasten M, Schaake S, Petkovic S, et al. Genotype-phenotype relations for the Parkinson's disease genes SNCA, LRRK2, VPS35: MDSGene systematic review. Mov Disord 2018; 33(12): 1857-70.

Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 2010; 38(16): e164.

Wickremaratchi MM, Knipe MD, Sastry BS, Morgan E, Jones A, Salmon R, et al. The motor phenotype of Parkinson's disease in relation to age at onset. Mov Disord 2011; 26(3): 457-63.

Wilson GR, Sim JC, McLean C, Giannandrea M, Galea CA, Riseley JR, et al. Mutations in RAB39B cause X-linked intellectual disability and early-onset Parkinson disease with alpha-synuclein pathology. Am J Hum Genet 2014; 95(6): 729-35.

Wullner U, Kaut O, deBoni L, Piston D, Schmitt I. DNA methylation in Parkinson's disease. J Neurochem 2016; 139 Suppl 1: 108-20.

Xu Q, Li K, Sun Q, Ding D, Zhao Y, Yang N, et al. Rare GCH1 heterozygous variants contributing to Parkinson's disease. Brain 2017; 140(7): e41.

Yan W, Tang B, Zhou X, Lei L, Li K, Sun Q, et al. TMEM230 mutation analysis in Parkinson's disease in a Chinese population. Neurobiol Aging 2017; 49: 219 e1- e3.

Yang JO, Kim WY, Jeong SY, Oh JH, Jho S, Bhak J, et al. PDbase: a database of Parkinson's disease-related genes and genetic variation using substantia nigra ESTs. BMC Genomics 2009; 10 Suppl 3: S32.

Yang N, Zhao Y, Liu Z, Zhang R, He Y, Zhou Y, et al. Systematically analyzing rare variants of autosomal-dominant genes for sporadic Parkinson's disease in a Chinese cohort. Neurobiol Aging 2019; 76: 215 e1- e7.

Additional file 1: Table S1. Summary of collected data in present study.

Additional file 2: Table S2. Weight scheme of various genetic data for prioritizing candidate genes.

Additional file 3: Table S3. Data source included in Gene4PD.

Additional file 4: Table S4. Gene locus and disease-causing genes of Parkinson disease.

Additional file 5: Table S5. The biological progress information of gene ontology (GO) terms involved into the genes in PPI network.

Additional file 6: Fig. S1. The interconnectivities among PD-associated genes accessed by permutation tests. (a-b) The interconnectivities among high confidence/strong PD-associated genes. (c-d) The interconnectivities among suggestive associated genes. (e-f) The connection between high confidence/strong PD-associated genes and suggestive associated genes.

Additional file 7: Fig. S2. Snapshot of gene-level implications in Gene4PD. Six sections, including ‘basic information’, ‘gene functions’, ‘phenotype and disease’, ‘gene expression’, ‘Variants in different population’, ‘Drug-gene interaction’ are illustrated in gene-level implications.

Additional file 8: Fig. S3. Snapshot of analysis panel in Gene4PD. Two main sections, including inputting cosegregated variants panel, and choosing specify annotation datasets were illustrated in the analysis panel.

Supplementaldata.docx

Download PDF

Journal Publication

published 26 Apr, 2021

Read the published version in Frontiers in Neuroscience →

Version 1

posted

You are reading this latest preprint version

Gene4PD: a comprehensive genetic database of Parkinson's disease

Status:

Journal Publication

Version 1

Abstract

Figures

Background

Methods

Data collection

Data annotation and PAG prioritization

The verification of functional correlation and PAG network

Association between age at onset and PAGs

Database construction and interface

Results

Genomic data integration and PAG prioritization

PAGs were functionally correlated

AAO is associated with multiple genetic components

Discussion

Conclusion

Abbreviations

Declarations

References

Supplemental Information

Supplementary Files

Status:

Journal Publication

Version 1