Building an Interoperable Rare Disease Multi-omic Resource: The GREGoR Data Model and Dataset

The GREGoR Consortium developed a modular, interoperable data model to standardize multi-omic and phenotypic data across diverse clinical sites, successfully harmonizing information from over 12,000 participants to enable large-scale rare disease research and cross-study analysis.

The Gist

Imagine trying to solve a massive, global puzzle where every piece comes from a different factory. Some pieces are shaped like squares, others like triangles, and they all speak different languages. This is what researchers faced when trying to study rare diseases. They had genetic clues and patient stories coming from hospitals and labs all over the world, but because everyone used their own unique way of writing things down, the pieces didn't fit together. You couldn't easily compare one patient's story with another's, making it hard to find the patterns needed to solve the mystery of these diseases.

To fix this, a group called the GREGoR Consortium decided to build a universal "instruction manual" for how to write down this information. Think of this manual as a standardized LEGO set. Before, every scientist was building with their own custom blocks that only fit their own creations. Now, they all use the same set of blocks with the same connectors.

Here is how their new system, called the GREGoR Data Model, works in everyday terms:

• The Modular Design: Imagine a filing cabinet where you can easily swap out drawers. This system is built like that. It allows researchers to plug in different types of "omics" data (which are like different layers of biological information, such as DNA, proteins, or cell functions) and still keep everything neatly organized under one person's file.
• The "Who Did It" Tag: A key feature of this system is like a label on a recipe. If a scientist discovers a genetic clue using a specific high-tech machine, the system tags that clue with exactly which machine found it. This ensures that if the technology changes later, we still know exactly where the original discovery came from.
• Connecting the Dots: Because everyone is now using the same "LEGO blocks," the researchers were able to snap together data from 12,292 individuals across 5,029 families. This created a single, giant, harmonized dataset that is ready to be analyzed, rather than a messy pile of incompatible files.

The paper claims that by using this flexible and collaborative "instruction manual," the consortium has successfully created a massive, shared resource. They are now sharing this data publicly, and other groups working on rare diseases are starting to use this same manual. The result is that different research teams can finally talk to each other and combine their findings, making the work of solving rare diseases much faster and more effective.

Technical Summary

Technical Summary: Building an Interoperable Rare Disease Multi-omic Resource

Problem Statement
Rare disease research and clinical diagnosis fundamentally depend on integrating genomic and phenotypic data collected across diverse clinical sites. However, the field has been hindered by a lack of widely adopted standards for representing genomic data and associated metadata. This absence of standardization has severely limited data interoperability, the ability to reuse data, and the capacity for cross-study analysis, creating barriers to effective collaboration and discovery.

Methodology
To address these challenges, the Genomics Research to Elucidate the Genetics of Rare Diseases (GREGoR) Consortium was established with the dual goals of investigating challenging rare disease cases and evaluating emerging multi-omic technologies for clinical translation. Central to this effort was the development of a common Consortium Data Model, created in partnership with domain experts.

The methodology focused on standardizing the capture of metadata across four distinct levels:

1. Participant-level
2. Family-level
3. Phenotype-level
4. Assay-level

A critical architectural feature of this model is its modularity. This design specifically supports the linking of multiple data versions derived from various omic technologies to a single individual. Furthermore, the model ensures the precise attribution of genetic findings to the specific technology used for their initial discovery, maintaining data provenance and integrity.

Key Contributions
The primary contribution of this work is the successful adoption and implementation of the GREGoR Data Model, which has enabled the generation and public release of a harmonized, analysis-ready Consortium Dataset. The model serves as a standardized framework that allows distributed research sites to coordinate data integration effectively. Additionally, the work demonstrates the practical application of this model in managing complex, multi-omic datasets at scale.

Results
The application of the GREGoR Data Model has yielded a substantial, publicly available resource. The most recent release of the Consortium Dataset includes:

• 12,292 participants
• 5,029 families
• Comprehensive phenotypic, family, and multi-omic data

Beyond the immediate dataset, the paper reports that other rare disease data sharing efforts are beginning to adopt this specific data model, indicating its potential utility beyond the initial consortium.

Significance
The paper posits that the GREGoR initiative demonstrates that a collaborative, flexible, and scalable data model is capable of enabling large-scale rare disease research. By facilitating cross-center data harmonization, the model directly addresses the historical issues of interoperability. The adoption of this framework is claimed to empower rare disease research by enabling cross-consortium analyses and providing a robust foundation for the integration of complex multi-omic data.