1 of 19

LinkML in Rare Diseases: Ontology-Based Interoperability for Clinical Registries and Analysis

Adam Graefe

LinkML community call

Thursday, 16th of January 2026

2 of 19

2

F A I R

indable

iccessible

interoperable

eusable

LinkML community call – 15th of January 2026 | Adam Graefe

3 of 19

Interoperability

  • Semantic Interoperability

Shared and precise interpretation of medical data

3

  • Syntactic Interoperability

Communication via compatible formats and protocols

LinkML community call – 15th of January 2026 | Adam Graefe

4 of 19

4

MII

IPS

FHIR

Genomics Reporting

EHDS

Patient

Observation

Condition

Specimen

Location

Practitioner

Procedure

Device

Encounter

Episode of Care

Account

Health Level 7

Fast

Healthcare

Interoperability

Resources

Hospital Information Systems

int. RD-CDM

LinkML community call – 15th of January 2026 | Adam Graefe

5 of 19

5

FHIR

Genomics Reporting

EHDS

IPS

MII

Abaza H et al .Domain- Specific Common Data Elements for Rare Disease Registration: Conceptual approach of a European Joint Initiative toward Semantic Interoperability in Rare Disease research. JMIR Medical Informatics. 2022 May 20;10(5):e32158. https://doi.org/10.2196/32158

int. RD-CDM

LinkML community call – 15th of January 2026 | Adam Graefe

6 of 19

6

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

https://www.nature.com/articles/s41587-022-01357-4

LinkML community call – 15th of January 2026 | Adam Graefe

7 of 19

7

IPS

EHDS

IPS

Genomics Reporting

MII

________

Analysis

int. RD-CDM

LinkML community call – 15th of January 2026 | Adam Graefe

8 of 19

8

IPS

IPS

Genomics Reporting

MII

EHDS

https://github.com/BIH-CEI/rd-cdm

int. RD-CDM

Common Data Set

LinkML community call – 15th of January 2026 | Adam Graefe

9 of 19

9

data_elements.yaml

rd_cdm_v2_0_2.yaml

value_sets.yaml

code_systems.yaml

rd_cdm.yaml (root class)

>_ rd-cdm-merge

rd_cdm_v2_0_2.json

rd_cdm_v2_0_2.csv

>_ rd-cdm-json

(LinkML loaders)

>_ rd-cdm-csv

rd_cdm_pydantic.py

rd_cdm.py

LinkML runtime

LinkML pydantic gen.

>_ rd-cdm-validate

LinkML community call – 15th of January 2026 | Adam Graefe

10 of 19

10

IPS

IPS

Genomics Reporting

MII

EHDS

int. RD-CDM

LinkML community call – 15th of January 2026 | Adam Graefe

11 of 19

11

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

LinkML community call – 15th of January 2026 | Adam Graefe

12 of 19

12

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

RareLink-CDM

LinkML schema

REDCap instruments

_____________

Local REDCap site

________________

FHIR resources

Hospital Information System

________

Registry Linkage

________

Analysis

________________

Phenopackets

_____________

A local rare disease �cohort / registry site

Prospective RD data

GenomicsReporting

R4

2.0

Manual Data Capture Guide

Semi-Automated �Import (LinkML Map)

CDIS Module

toFHIR �Module

RareLink- �Phenopackets Module

Retrospective (tabular) data

RD-CDM

LinkML community call – 15th of January 2026 | Adam Graefe

13 of 19

13

RareLink- Phenopackets engine

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

RareLink-CDM

LinkML schema

REDCap instruments

REDCap data �dictionary

LinkML schema definitions

REDCap instruments�(variables & choices�w/ embedded ontologies)

Meta schema

Code systems

Repeated elements

Custom data �models�& disease-�specific �extensions

LinkML�validate

Corresponding �Pydantic &

Python classes

RD-CDM

Raw REDCap �cohort data

________________

Phenopackets

14 of 19

14

LinkML community call – 15th of January 2026 | Adam Graefe

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

15 of 19

15

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

https://rarelink.readthedocs.io/

LinkML community call – 15th of January 2026 | Adam Graefe

16 of 19

16

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

Undiagnosed �Disease �Programme

LinkML community call – 15th of January 2026 | Adam Graefe

17 of 19

17

IPS

EHDS

IPS

Genomics Reporting

MII

int. RD-CDM

LinkML community call – 15th of January 2026 | Adam Graefe

18 of 19

Ideas and next steps

  • Improved automation for RD-CDM and RareLink
  • LinkML-based rare disease templates for new/existing data models and cohorts
  • Ongoing and new collaborations with clinical RD registries globally
  • Exchange and alignment with international entities (HL7, GA4GH, GDI, JARDIN, etc.)
  • Better integration and use of further LinkML tools and OAK
  • Deep integration of agentic LLM tools (i.e. ai4curation, etc.) for REDCap�

🡪 Open-source and community-based: �we’re happy to collaborate and look forward to feedback!

18

LinkML community call – 15th of January 2026 | Adam Graefe

19 of 19

Thank you very much!����

adam.graefe@charite.de

LinkML community call – 15th of January 2026 | Adam Graefe