Novel Framework Enhances Rare Disease Gene Discovery
Publication Title: PERADIGM: Phenotype embedding similarity-based rare disease gene mapping
Summary
- Question
- This study introduced PERADIGM, a computational framework designed to identify genes associated with rare diseases by analyzing patient phenotype similarities derived from electronic health records (EHRs). Researchers aimed to improve rare disease gene mapping by incorporating richer phenotype data and leveraging natural language processing techniques.
- Why it Matters
- Rare diseases often affect small populations, making it difficult to identify genetic contributors using traditional methods that rely on binary disease classifications. By analyzing comprehensive medical histories, PERADIGM enhances our ability to uncover genes linked to rare diseases and related phenotypes. This approach could lead to better understanding of genetic mechanisms, improved diagnostic precision, and more targeted therapies, benefiting both the scientific community and patients with rare conditions.
- Methods
- The researchers applied PERADIGM to analyze data from the UK Biobank, which includes genetic and EHR data for over 500,000 participants. They focused on three rare diseases: autosomal dominant polycystic kidney disease (ADPKD), Marfan syndrome, and neurofibromatosis type 1 (NF1). Using an embedding model, PERADIGM calculated patient similarity scores based on ICD-10 codes and compared phenotypic profiles of individuals diagnosed with these diseases to those carrying rare genetic variants.
- Key Findings
- PERADIGM successfully identified well-known disease-associated genes (e.g., PKD1 and PKD2 for ADPKD, FBN1 for Marfan syndrome, and NF1 for NF1) and revealed additional candidate genes, such as IFT140 for ADPKD and COL5A1 for Marfan syndrome. These genes may influence disease manifestations or modify phenotypes. The framework demonstrated improved sensitivity compared to traditional methods by integrating nuanced phenotype information, enabling the detection of broader genotype-phenotype relationships.
- Implications
- The findings suggest that PERADIGM is a powerful tool for identifying genes associated with rare diseases, offering deeper insights into their genetic architecture and clinical variability. This approach could help refine diagnostic criteria, uncover genetic modifiers, and identify potential therapeutic targets. By incorporating rich phenotype data, PERADIGM addresses limitations of traditional methods, advancing precision medicine and rare disease research.
- Next Steps
- The authors recommend expanding PERADIGM to include larger and more diverse datasets to enhance its applicability across populations. Future research should also explore integrating advanced embedding models, such as BERT, to capture longitudinal phenotype data and refine disease-specific analyses further.
- Funding Information
- This research was supported by the National Institute of Child Health and Human Development (NICHD) and the National Institute of General Medical Sciences (NIGMS) under grants R03 HD100883 and R01 GM134005. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Yale University also provided funding and support for this research.
Full Citation
Zheng W, Xie Y, Gu J, Li H, Somlo S, Besse W, Zhao H. PERADIGM: Phenotype embedding similarity-based rare disease gene mapping. PLOS Genetics 2025, 21: e1011976. PMID: 41411243, PMCID: PMC12714201, DOI: 10.1371/journal.pgen.1011976.
This AI-assisted summary has been reviewed and approved by at least one of the study's authors to ensure it accurately reflects the research.
Authors
Wangjie Zheng
First AuthorHongyu Zhao, PhD
Last AuthorIra V. Hiscock Professor of Biostatistics, Professor of Genetics and Professor of Statistics and Data Science
Additional Yale School of Medicine Authors
Other Authors
Research Themes
Keywords
Concepts
- Rare disease genetics;
- Disease gene discovery;
- Autosomal dominant polycystic kidney disease;
- Comprehensive phenotypic information;
- Disease genetics;
- Gene discovery;
- Gene mapping;
- Phenotypic information;
- Individual phenotypes;
- Electronic health records;
- Binary disease status;
- Disease-specific phenotypes;
- Dominant polycystic kidney disease;
- Genes;
- Phenotype;
- Polycystic kidney disease;
- UK Biobank dataset;
- Neurofibromatosis type 1;
- Precision medicine;
- Health records;
- Biobank dataset;
- ICD-10;
- Association method;
- Statistical power;
- Rare disease