Phenome-wide identification of therapeutic genetic targets, leveraging knowledge graphs, graph neural networks, and UK Biobank data

Middleton, Lawrence; Melas, Ioannis; Vasavda, Chirag; Raies, Arwa; Rozemberczki, Benedek; Dhindsa, Ryan S.; Dhindsa, Justin S.; Weido, Blake; Wang, Quanli; Harper, Andrew R.; Edwards, Gavin; Petrovski, Slavé; Vitsios, Dimitrios

doi:10.1126/sciadv.adj1424

Cited by 2 publications

(1 citation statement)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We next tested whether there was orthogonal evidence to support ITSN1's association with PD using Mantis-ML. 17 This machine-learning framework predicts genotype-phenotype relationships using a vast set of features, including gene expression data, genetic intolerance, PPI networks, and a knowledge graph with over 8.7M edges. Remarkably, across 2,575 studied phenotypes, "hand tremor", "inability to walk", and "slurred speech" were among the top 10 Mantis-ML predictions for ITSN1-related phenotypes (Fig.…”

Section: Supporting Evidence For Itsn1 As a Pd Risk Genementioning

confidence: 99%

Haploinsufficiency ofITSN1is associated with Parkinson’s disease

Spargo,

Sands,

Juan

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Background Despite its significant heritability, the genetic underpinnings of Parkinson disease (PD) remain incompletely understood, particularly the role of rare variants. Advances in population-scale sequencing now provide an unprecedented opportunity to uncover additional large-effect rare genetic risk factors and expand our understanding of disease mechanisms. Methods We leveraged whole-genome sequence data with linked electronic health records from 490,560 UK Biobank participants, identifying 3,809 PD cases and 247,101 controls without a neurological disorder. We performed both variant- and gene-level association analyses to identify novel genetic associations with PD. We analyzed two additional independent case-control cohorts for replication (totaling 3,739 cases and 58,156 controls). Additionally, we performed functional validation of a novel PD association in a human synuclein-expressing Drosophila model. Findings In the UK Biobank, we replicated associations in well-established loci including GBA1 and LRRK2. We also identified a novel association between protein-truncating variants (PTVs) in ITSN1 and an increased risk of PD, with an effect size exceeding those of established loci (Fisher's Exact Test: p=6.1x10-7; Odds ratio [95% confidence interval] = 10.53 [5.20, 21.34]). We replicated the ITSN1 risk signal in a meta-analysis across all cohorts (Cochran-Mantel-Haenszel test p=5.7x10-9; Odds ratio [95% confidence interval] = 9.20 [4.66, 16.70]). In Drosophila, haploinsufficiency of the ITSN1 ortholog (Dap160) exacerbated α-synuclein-induced compound eye degeneration and motor deficits. Interpretation We establish ITSN1 as a novel risk gene for PD, with PTVs substantially increasing disease risk. ITSN1 encodes a scaffold protein involved in synaptic vesicle endocytosis, a critical pathway increasingly recognized in PD pathogenesis. Our findings highlight the power of large-scale sequencing coupled with preclinical functional modeling to identify rare variant associations and elucidate disease mechanisms.

show abstract

Section: Supporting Evidence For Itsn1 As a Pd Risk Genementioning

confidence: 99%

Haploinsufficiency ofITSN1is associated with Parkinson’s disease

Spargo,

Sands,

Juan

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Disease prediction with multi-omics and biomarkers empowers case–control genetic discoveries in the UK Biobank

Garg,

Karpinski,

Matelska

et al. 2024

Nat Genet

View full text Add to dashboard Cite

The emergence of biobank-level datasets offers new opportunities to discover novel biomarkers and develop predictive algorithms for human disease. Here, we present an ensemble machine-learning framework (machine learning with phenotype associations, MILTON) utilizing a range of biomarkers to predict 3,213 diseases in the UK Biobank. Leveraging the UK Biobank’s longitudinal health record data, MILTON predicts incident disease cases undiagnosed at time of recruitment, largely outperforming available polygenic risk scores. We further demonstrate the utility of MILTON in augmenting genetic association analyses in a phenome-wide association study of 484,230 genome-sequenced samples, along with 46,327 samples with matched plasma proteomics data. This resulted in improved signals for 88 known (P < 1 × 10−8) gene–disease relationships alongside 182 gene–disease relationships that did not achieve genome-wide significance in the nonaugmented baseline cohorts. We validated these discoveries in the FinnGen biobank alongside two orthogonal machine-learning methods built for gene–disease prioritization. All extracted gene–disease associations and incident disease predictive biomarkers are publicly available (http://milton.public.cgr.astrazeneca.com).

show abstract

Phenome-wide identification of therapeutic genetic targets, leveraging knowledge graphs, graph neural networks, and UK Biobank data

Cited by 2 publications

References 46 publications

Haploinsufficiency ofITSN1is associated with Parkinson’s disease

Haploinsufficiency ofITSN1is associated with Parkinson’s disease

Disease prediction with multi-omics and biomarkers empowers case–control genetic discoveries in the UK Biobank

Contact Info

Product

Resources

About