Biology subjects

Hardy, J.

Publications and source records attributed to Hardy, J..

At least 73 records · Page 4Linked to original sources

bioRxiv2019

An integrated genomic approach to dissect the genetic landscape regulating the cell-to-cell transfer of a-synuclein

Neuropathological and experimental evidence suggests that the cell-to-cell transfer of a-synuclein has an important role in the pathogenesis of Parkinsons disease (PD). However, the mechanism underlying this phenomenon is not fully understood. We undertook an siRNA, genome-wide high throughput screen to identify genes regulating the cell-to-cell transfer of a-synuclein. We transiently transfected HEK cells stably overexpressing a-synuclein with a construct encoding GFP-2a-aSynuclein-RFP. The cells expressing a-synuclein-RFP through transfection were double positive for GFP and RFP fluorescence, whereas the cells receiving it through transfer were positive only for RFP fluorescence. The amount of a-synuclein transfer was quantified by high content microscopy. A series of unbiased screens confirmed the involvement of 38 genes in the regulation of a-synuclein-RFP transfer. One of those hits was ITGA8, a candidate gene recently identified through a large PD genome wide association study (GWAS). Weighted gene co-expression network analysis (WGCNA) and weighted protein-protein network interaction analysis (WPPNIA) showed that the hits clustered in networks that included known PD Mendelian and GWAS risk genes more frequently than expected than random chance. Given the genetic overlap between a-synuclein transfer and PD, those findings provide supporting evidence for the importance of the cell-to-cell transfer of a-synuclein in the pathogenesis of PD, and expand our understanding of the mechanism of a-synuclein spread.

neuroscience↗

bioRxiv2019

PINOT: An Intuitive Resource for Integrating Protein-Protein Interactions

The past decade has seen the rise of omics data, for the understanding of biological systems in health and disease. This wealth of data includes protein-protein interaction (PPI) derived from both low and high-throughput assays, which is curated into multiple databases that capture the extent of available information from the peer-reviewed literature. Although these curation efforts are extremely useful, reliably downloading and integrating PPI data from the variety of available repositories is challenging and time consuming.\n\nWe here present a novel user-friendly web-resource called PINOT (Protein Interaction Network Online Tool; available at http://www.reading.ac.uk/bioinf/PINOT/PINOT_form.html) to optimise the collection and processing of PPI data from the IMEx consortium associated repositories (members and observers) and from WormBase for constructing, respectively, human and C. elegans PPI networks.\n\nUsers submit a query containing a list of proteins of interest for which PINOT will mine PPIs. PPI data is downloaded, merged, quality checked, and confidence scored based on the number of distinct methods and publications in which each interaction has been reported. Examples of PINOT applications are provided to highlight the performance, the ease of use and the potential applications of this tool.\n\nPINOT is a tool that allows users to survey the literature, extracting PPI data for a list of proteins of interest. The comparison with analogous tools showed that PINOT was able to extract similar numbers of PPIs while incorporating a set of innovative features. PINOT processes both small and large queries, it downloads PPIs live through PSICQUIC and it applies quality control filters on the downloaded PPI annotations (i.e. removing the need of manual inspection by the user). PINOT provides the user with information on detection methods and publication history for each of the downloaded interaction data entry and provides results in a table format that can be easily further customised and/or directly uploaded in a network visualization software.

bioinformatics↗

bioRxiv2019

A locked immunometabolic switch underlies TREM2 R47H loss of function in human iPSC--derived microglia

Loss-of-function genetic variants of triggering receptor expressed on myeloid cells 2 (TREM2) are linked with an enhanced risk of developing dementias. Microglia, the resident immune cell of the brain, express TREM2 and microglial responses are implicated in dementia pathways. In a normal surveillance state, microglia use oxidative phosphorylation for their energy supply, but rely on the ability to undergo a metabolic switch to glycolysis to allow them to perform rapid plastic responses. We investigated the role of TREM2 on microglial metabolic function in human patient iPSC-derived-microglia expressing loss of function variants in TREM2. We show that these TREM2 variant iPSC-microglia, including the Alzheimers disease R47H risk variant, exhibit significant metabolic deficits including a reduced mitochondrial respiratory capacity and an inability to perform a glycolytic immunometabolic switch. We determined that dysregulated PPAR{gamma}/p38MAPK signalling underlies the observed phenotypic deficits in TREM2 variants and that activation of these pathways can ameliorate the metabolic deficit in these cells and consequently rescue critical microglial cellular function such as {beta}-Amyloid phagocytosis. These findings have ramifications for microglial focussed-treatments in AD.

neuroscience↗

bioRxiv2019

A genome-wide genetic pleiotropy approach identified shared loci between multiple system atrophy and inflammatory bowel disease

We aimed to identify shared genetic background between multiple system atrophy (MSA) and autoimmune diseases by using the conjFDR approach. Our study showed significant genetic overlap between MSA and inflammatory bowel disease and identified DENND1B, C7, and RSP04 loci, which are linked to significant changes in methylation or expression levels of adjacent genes. We obtained evidence of enriched heritability involving immune/digestive categories. Finally, an MSA mouse model showed dysregulation of the C7 gene in the degenerating midbrain compared to wildtype mice. The results identify novel molecular mechanisms and implicate immune and gut dysfunction in MSA pathophysiology.

genetics↗

bioRxiv2019

Analysis of DNM3 and VAMP4 as genetic modifiers of LRRK2 Parkinsons disease

ObjectiveTo assess genetic modifiers of Parkinsons disease (PD) age at onset (AAO) penetrance in individuals carrying common and rare LRRK2 risk alleles\n\nMethodsWe analysed reported genetic modifier DNM3 rs2421947 in 724 LRRK2 p.G2019S heterozygotes using linear regression of AAO. We meta-analysed our data with previously published data (n=754). VAMP4 is in close proximity to DNM3 and is associated with PD. We analysed the effect of the rs11578699 VAMP4 variant on pG2019S penetrance in 786 LRRK2 p.G2019S heterozygotes. We also evaluated the impact of VAMP4 variants using AAO regression in 4882 patients with PD carrying a common LRRK2 risk variant (rs10878226).\n\nResultsThere was no evidence for linkage disequilibrium between DNM3 rs2421947 and VAMP4 rs11578699. Our linear regression AAO of 724 p.G2019S carriers showed no relationship between DNM3 rs2421947 and AAO (beta = -1.19, p = 0.55, n =708). Meta-analysis with previously published data did not indicate a significant effect on AAO (beta = -2.21, p = 0.083, n = 1304), but there was significant heterogeneity in the analyses of new and previously published data. VAMP4 rs11578699 was nominally associated with AAO in patients dichotomized by the common LRRK2 risk variant rs10878226 (beta=1.68, se=0.81 p=0.037).\n\nInterpretationAnalysis of DNM3 in previously unpublished data does not show an interaction between DNM3 and LRRK2 G2019S for AAO, however the inter-study heterogeneity may indicate ethnic-specific effects of DNM3 rs2421947. Analysis of sporadic PD patients stratified by the PD risk variant rs10878226 indicates a possible interaction between LRRK2 and VAMP4.

genetics↗

bioRxiv2019

Integration of eQTL and Parkinson’s disease GWAS data implicates 11 disease genes

Substantial genome-wide association study (GWAS) work in Parkinsons disease (PD) has led to an increasing number of loci shown reliably and robustly to be associated with the increased risk of the disease. Prioritising causative genes and pathways from these studies has proven problematic. Here, we present a comprehensive analysis of PD GWAS data with expression and methylation quantitative trait loci (eQTL/mQTL) using Colocalisation analysis (Coloc) and transcriptome-wide association analysis (TWAS) to uncover putative gene expression and splicing mechanisms driving PD GWAS signals. Candidate genes were further characterised by determining cell-type specificity, weighted gene co-expression (WGNCA) and protein-protein interaction (PPI) networks.\n\nGene-level analysis of expression revealed 5 genes (WDR6, CD38, GPNMB, RAB29, TMEM163) that replicated using both Coloc and TWAS analyses in both GTEx and Braineac expression datasets. A further 6 genes (ZRANB3, PCGF3, NEK1, NUPL2, GALC, CTSB) showed evidence of disease-associated splicing effects. Cell-type specificity analysis revealed that gene expression was overall more prevalent in glial cell-types compared to neurons. The WGNCA analysis showed that NUPL2 is a key gene in 3 modules implicated in catabolic processes related with protein ubiquitination (protein ubiquitination (p=7.47e-10) and ubiquitin-dependent protein catabolic process (p = 2.57e-17) in nucleus accumbens, caudate and putamen, while TMEM163 and ZRANB3 were both important in modules indicating regulation of signalling (p=1.33e-65] and cell communication (p=7.55e-35) in the frontal cortex and caudate respectively. PPI analysis and simulations using random networks demonstrated that the candidate genes interact significantly more with known Mendelian PD and parkinsonism proteins than would be expected by chance. The proteins core proteins this network were enriched for regulation of the ERBB receptor tyrosine protein kinase signalling pathways.\n\nTogether, these results point to a number of candidate genes and pathways that are driving the associations observed in PD GWAS studies.

neuroscience↗

bioRxiv2019

The Parkinsons Disease Mendelian Randomization Research Portal

BackgroundMendelian randomization (MR) is a method for exploring observational associations to find evidence of causality.\n\nObjectiveTo apply MR between multiple risk factors/phenotypic traits (exposures) and Parkinsons disease (PD) in a large, unbiased manner, and to create a public resource for research.\n\nMethodsWe used two-sample MR in which the summary statistics relating to SNPs from genome wide association studies (GWASes) of 5,839 exposures curated on MR Base were used to assess causal relationships with PD. We selected the highest quality exposure GWASes for this report (n=401). For the disease outcome, summary statistics from the largest published PD GWAS were used. For each exposure, the causal effect on PD was assessed using the inverse variance weighted (IVW) method, followed by a range of sensitivity analyses. We used a false discovery rate (FDR) corrected p-value of <0.05 from the IVW analysis to prioritize traits of interest.\n\nResultsWe observed evidence for causal associations between twelve exposures and risk of PD. Of these, nine were causal effects related to increasing adiposity and decreasing risk of PD. The remaining top exposures that affected PD risk were tea drinking, time spent watching television and forced vital capacity, but the latter two appeared to be biased by violations of underlying MR assumptions.\n\nDiscussionWe present a new platform which offers MR analyses for a total of 5,839 GWASes versus the largest PD GWASes available (https://pdgenetics.shinyapps.io/pdgenetics/). Alongside, we report further evidence to support a causal role for adiposity on lowering the risk of PD.

neuroscience↗

bioRxiv2019

ERASE: Extended Randomization for assessment of annotation enrichment in ASE datasets

Genome-wide association studies (GWAS) have identified thousands of genetic variants associated with various human phenotypes and many of these loci are thought to act at a molecular level by regulating gene expression. Detection of allele specific expression (ASE), namely preferential usage of an allele at a transcribed locus, is an increasingly important means of studying the genetic regulation of gene expression. However, there are currently a paucity of tools available to link ASE sites with GWAS risk loci. Existing integration methods first use ASE sites to infer cis-acting expression quantitative trait loci (eQTL) and then apply eQTL-based approaches. ERASE is a method that assesses the enrichment of risk loci amongst ASE sites directly. Furthermore, ERASE enables additional biological insights to be made through the addition of other SNP level annotations. ERASE is based on a randomization approach and controls for read depth, a significant confounder in ASE analyses. In this paper, we demonstrate that ERASE can efficiently detect the enrichment of eQTLs and risk loci within ASE data and that it remains sensitive even when used with underpowered GWAS datasets. Finally, using ERASE in combination with GWAS data for Parkinsons disease and data on the splicing potential of individual SNPs, we provide evidence to suggest that risk loci for Parkinsons disease are enriched amongst ASEs likely to affect splicing. Thus, we show that ERASE is an important new tool for the integration of ASE and GWAS data, capable of providing novel insights into the pathophysiology of complex diseases.

bioinformatics↗

bioRxiv2018

Incomplete annotation of OMIM genes is likely to be limiting the diagnostic yield of genetic testing, particularly for neurogenetic disorders

Although the increasing use of whole-exome and whole-genome sequencing have improved the yield of genetic testing for Mendelian disorders, an estimated 50% of patients still leave the clinic without a genetic diagnosis. This can be attributed in part to our lack of ability to accurately interpret the genetic variation detected through next-generation sequencing. Variant interpretation is fundamentally reliant on accurate and complete gene annotation, however numerous reports and discrepancies between gene annotation databases reveals that the knowledge of gene annotation remains far from comprehensive. Here, we detect and validate transcription in an annotation-agnostic manner across all 41 different GTEx tissues, then connect novel transcription to known genes, ultimately improving the annotation of 63% of the known OMIM-morbid genes. We find the majority of novel transcription to be tissue-specific in origin, with brain tissues being most susceptible to misannotation. Furthermore, we find that novel transcribed regions tend to be poorly conserved, but are significantly depleted for genetic variation within humans, suggesting they are functionally significant and potentially have human-specific functions. We present our findings through an online platform vizER, which enables individual genes to be visualised and queried for evidence of misannotation. We also release all tissue-specific transcriptomes in a BED format for ease of integration with whole-genome sequencing data. We anticipate that these resources will improve the diagnostic yield for a wide range of Mendelian disorders.

bioinformatics↗

bioRxiv2018

Trem2 promotes anti-inflammatory responses in microglia and is suppressed under pro-inflammatory conditions

Genome-wide association studies have reported that, amongst other microglial genes, variants in TREM2 can profoundly increase the incidence of developing Alzheimers disease (AD). We have investigated the role of TREM2 in primary microglial cultures from wild type mice by using siRNA to decrease Trem2 expression, and in parallel from knock-in mice heterozygous or homozygous for the Trem2 R47H AD risk variant. The prevailing phenotype of Trem2 R47H knock-in mice was decreased expression levels of Trem2 in microglia, which resulted in decreased density of microglia in the hippocampus. Overall, primary microglia with reduced Trem2 expression, either by siRNA or from the R47H knock-in mice, displayed a similar phenotype. Comparison of the effects of decreased Trem2 expression under conditions of LPS pro-inflammatory or IL-4 anti-inflammatory stimulation revealed the importance of Trem2 in driving a number of the genes up-regulated in the anti-inflammatory phenotype. RNA-seq analysis showed that IL-4 induced the expression of a programme of genes including Arg1 and Ap1b1 in microglia, which showed an attenuated response to IL-4 when Trem2 expression was decreased. Genes showing a similar expression profile to Arg1 were enriched for STAT6 transcription factor recognition elements in their promoter, and Trem2 knockdown decreased levels of the transcription factor STAT6. LPS-induced pro-inflammatory stimulation suppressed Trem2 expression, thus preventing TREM2s anti-inflammatory drive. Given that anti-inflammatory signaling is associated with tissue repair, understanding the signaling mechanisms downstream of Trem2 in coordinating the pro- and anti-inflammatory balance of microglia, particularly mediating effects of the IL-4-regulated anti-inflammatory pathway, has important implications for fighting neurodegenerative disease. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=189 HEIGHT=200 SRC="FIGDIR/small/449884v2_ufig1.gif" ALT="Figure 1"> View larger version (31K): org.highwire.dtl.DTLVardef@1393670org.highwire.dtl.DTLVardef@1251691org.highwire.dtl.DTLVardef@1eca1c5org.highwire.dtl.DTLVardef@70a30a_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2018

Moving beyond neurons: the role of cell type-specific gene regulation in Parkinson’s disease heritability

Parkinsons disease (PD), with its characteristic loss of nigrostriatal dopaminergic neurons and deposition of -synuclein in neurons, is often considered a neuronal disorder. However, in recent years substantial evidence has emerged to implicate glial cell types, such as astrocytes and microglia. In this study, we used stratified LD score regression and expression-weighted cell-type enrichment together with several brain-related and cell-type-specific genomic annotations to connect human genomic PD findings to specific brain cell types. We found that PD heritability does not enrich in global and regional brain annotations or brain-related cell-type-specific annotations. Likewise, we found no enrichment of PD susceptibility genes in brain-related cell types. In contrast, we demonstrated a significant enrichment of PD heritability in a curated lysosomal gene set specifically expressed in astrocytic and microglial subtypes. Our results suggest that PD risk loci do not lie in specific cell types or individual brain regions, but rather in global cellular processes to which cell types may have varying vulnerability.

genomics↗

bioRxiv2018

Parkinson disease age of onset GWAS: defining heritability, genetic loci and a-synuclein mechanisms

Increasing evidence supports an extensive and complex genetic contribution to Parkinsons disease (PD). Previous genome-wide association studies (GWAS) have shed light on the genetic basis of risk for this disease. However, the genetic determinants of PD age of onset are largely unknown. Here we performed an age of onset GWAS based on 28,568 PD cases. We estimated that the heritability of PD age of onset due to common genetic variation was ~0.11, lower than the overall heritability of risk for PD (~0.27) likely in part because of the subjective nature of this measure. We found two genome-wide significant association signals, one at SNCA and the other a protein-coding variant in TMEM175, both of which are known PD risk loci and a Bonferroni corrected significant effect at other known PD risk loci, INPP5F/BAG3, FAM47E/SCARB2, and MCCC1. In addition, we identified that GBA coding variant carriers had an earlier age of onset compared to non-carriers. Notably, SNCA, TMEM175, SCARB2, BAG3 and GBA have all been shown to either directly influence alpha-synuclein aggregation or are implicated in alpha-synuclein aggregation pathways. Remarkably, other well-established PD risk loci such as GCH1, MAPT and RAB7L1/NUCKS1 (PARK16) did not show a significant effect on age of onset of PD. While for some loci, this may be a measure of power, this is clearly not the case for the MAPT locus; thus genetic variability at this locus influences whether but not when an individual develops disease. We believe this is an important mechanistic and therapeutic distinction. Furthermore, these data support a model in which alpha-synuclein and lysosomal mechanisms impact not only PD risk but also age of disease onset and highlights that therapies that target alpha-synuclein aggregation are more likely to be disease-modifying than therapies targeting other pathways.

genetics↗

bioRxiv2018

Genetic variability in response to Aβ deposition influences Alzheimer’s risk

Genetic analysis of late-onset Alzheimers disease risk has previously identified a network of largely microglial genes that form a transcriptional network. In transgenic mouse models of amyloid deposition we have previously shown that the expression of many of the mouse orthologs of these genes are co-ordinately up-regulated by amyloid deposition. Here we investigate whether systematic analysis of other members of this mouse amyloid-responsive network predicts other Alzheimers risk loci. This statistical comparison of the mouse amyloid-response network with Alzheimers disease genome-wide association studies identifies 5 other genetic risk loci for the disease (OAS1, CXCL10, LAPTM5, ITGAM and LILRB4). This work suggests that genetic variability in the microglial response to amyloid deposition is a major determinant for Alzheimers risk.\n\nOne Sentence SummaryIdentification of 5 new risk loci for Alzheimers by statistical comparison of mouse A{beta} microglial response with gene-based SNPs from human GWAS

neuroscience↗

bioRxiv2018

Gene-Based Analysis in HRC Imputed Genome Wide Association Data Identifies Three Novel Genes For Alzheimer’s Disease

A novel POLARIS gene-based analysis approach was employed to compute gene-based polygenic risk score (PRS) for all individuals in the latest HRC imputed GERAD (N cases=3,332 and N controls=9,832) data using the International Genomics of Alzheimers Project summary statistics (N cases=13,676 and N controls=27,322, excluding GERAD subjects) to identify the SNPs and weight their risk alleles for the PRS score. SNPs were assigned to known, protein coding genes using GENCODE (v19). SNPs are assigned using both 1) no window around the gene and 2) a window of 35kb upstream and 10kb downstream to include transcriptional regulatory elements. The overall association of a gene is determined using a logistic regression model, adjusting for population covariates.\n\nThree novel gene-wide significant genes were determined from the POLARIS gene-based analysis using a gene window; PPARGC1A, RORA and ZNF423. The ZNF423 gene resides in an Alzheimers disease (AD)-specific protein network which also includes other AD-related genes. The PPARGC1A gene has been linked to energy metabolism and the generation of amyloid beta plaques and the RORA has strong links with genes which are differentially expressed in the hippocampus. We also demonstrate no enrichment for genes in either loss of function intolerant or conserved noncoding sequence regions.

genetics↗

bioRxiv2018

Variation at the TRIM11 locus modifies Progressive Supranuclear Palsy phenotype

ObjectiveThe basis for clinical variation related to underlying Progressive Supranuclear Palsy (PSP) pathology is unknown. We performed a genome wide association study (GWAS) to identify genetic determinants of PSP phenotype.\n\nMethodsTwo independent pathological and clinically diagnosed PSP cohorts were genotyped and phenotyped to create Richardsons syndrome (RS) and non-RS groups. We carried out separate logistic regression GWAS to compare RS and non-RS groups and then combined datasets to carry out a whole cohort analysis (RS=367, non-RS=130). We validated our findings in a third cohort by referring to data from 100 deeply phenotyped cases from a recent GWAS. We assessed the expression/co-expression patterns of our identified genes and used our data to carry out gene-based association testing.\n\nResultsOur lead single nucleotide polymorphism (SNP), rs564309, showed an association signal in both cohorts, reaching genome wide significance in our whole cohort analysis - OR 5.5 (3.2-10.0), p-value 1.7x10-9. rs564309 is an intronic variant of the tripartite motif-containing protein 11 (TRIM11) gene, a component of the ubiquitin proteasome system (UPS). In our third cohort, minor allele frequencies of surrogate SNPs in high linkage disequilibrium with rs564309 replicated our findings. Gene based association testing confirmed an association signal at TRIM11. We found that TRIM11 is predominantly expressed neuronally, in the cerebellum and basal ganglia.\n\nInterpretationOur study suggests that the TRIM11 locus is a genetic modifier of PSP phenotype and potentially adds further evidence for the UPS having a key role in tau pathology, therefore representing a target for disease modifying therapies.

neuroscience↗

bioRxiv2018

Abnormal microtubule dynamics disrupt nucleocytoplasmic transport in tau-mediated frontotemporal dementia

The neuronal microtubule-associated protein tau (MAPT) is central to the pathogenesis of many dementias, including Alzheimers disease. Autosomal dominant mutations in MAPT cause inherited frontotemporal dementia (FTD), but the underlying pathogenic mechanisms are unclear. Using human stem cell models of FTD due to MAPT mutations, we find that tau becomes hyperphosphorylated and mislocalises to neuronal cell bodies and dendrites in cortical neurons, recapitulating a key early event in FTD. Mislocalised tau in the cell body leads to abnormal microtubule dynamics in FTD-MAPT neurons that grossly deform the nuclear membrane, resulting in defective nucleocytoplasmic transport. Neurons in the post-mortem human FTD-MAPT cortex have a high incidence of nuclear deformation, indicating that tau-mediated nuclear membrane dysfunction is an important pathogenic process in FTD. Defects in nucleocytoplasmic transport in FTD point to important commonalities in the pathogenic mechanisms of both tau-mediated dementias and ALS-FTD due to TDP-43 and C9orf72 mutations.

neuroscience↗

bioRxiv2018

Histone deposition plays an active role in recombination-dependent replication to balance genome2 stability upon replication stress

Replication stress poses a serious threat to genome and epigenome stability. Recombination-Dependent-Replication (RDR) ensures DNA synthesis resumption from arrested forks. Despite the identification of chromatin restoration pathways during DNA repair processes, crosstalk coupling RDR and chromatin assembly is largely unexplored. Here, we addressed the contribution of chromatin assembly to replication stress in fission yeast. We expressed a mutated histone (H3-H113D) to genetically impair replication-dependent chromatin assembly by destabilizing (H3-H4)2 tetramer. We established that DNA synthesis-dependent histone deposition is required for the completion of RDR. Histone deposition prevents joint-molecules from Rqh1-dependent disassembly, a crosstalk contributing to cell survival upon replication stress but channeling RDR towards deleterious events. Asf1 and CAF-1 act in RDR and CAF-1 recruitment to DNA synthesis associated to RDR requires the HR factor Rad52. Our data establish that CAF-1 counteracts Rqh1 activity at sites of replication stress by promoting repair synthesis-coupled histone deposition. These results demonstrate that histone deposition plays an active role in fine-tuning RDR, a benefit counterbalanced by stabilizing at-risk joint molecules for genome stability.

genetics↗

bioRxiv2018

G2P: Using machine learning to understand and predict genes causing rare neurological disorders

To facilitate precision medicine and neuroscience research, we developed a machine-learning technique that scores the likelihood that a gene, when mutated, will cause a neurological phenotype. We analysed 1126 genes relating to 25 subtypes of Mendelian neurological disease defined by Genomics England (March 2017) together with 154 gene-specific features capturing genetic variation, gene structure and tissue-specific expression and co-expression. We randomly re-sampled genes with no known disease association to develop bootstrapped decision-tree models, which were integrated to generate a decision tree-based ensemble for each disease subtype. Genes generating larger numbers of distinct transcripts and with higher probability of having missense mutations in normal individuals were significantly more likely to cause neurological diseases. Using mouse-mutant phenotypic data we tested the accuracy of gene-phenotype predictions and found that for 88% of all disease subtypes there was a significant enrichment of relevant phenotypic abnormalities when predicted genes were mutated in mice and in many cases mutations produced specific and matching phenotypes. Furthermore, using only newly identified genes included in the Genomics England November 2017 release, we assessed our gene-phenotype predictions and showed an 8.3 fold enrichment relative to chance for correct predictions. Thus, we demonstrate both the explanatory and predictive power of machine-learning-based models in neurological disease.

genetics↗