GA4GH Clinical & Phenotypic Data Capture & Exchange Work Stream

Minutes and Actions 2026-2027

Previous notes: 2023-2025 | 2021-2022 | 2019-2020 | 2017-2018

Learn more: GA4GH Work Stream page on the GA4GH website
Questions and feedback to beatrice.amos@ga4gh.org

Resources

GA4GH Slack Workspace: http://ga4gh.slack.com 

• Clin/Pheno channels: #clin-pheno_ws, #cp-pedigree, #cp-exposure-standards

Meeting protocols

• By participating in meetings, attendees agree to adhere to the GA4GH Standards for Professional Conduct.
• Meetings may be recorded for note-taking purposes. Recordings will be deleted within three months of the meeting taking place.

Resources        1

Agendas & Notes        2

2026-09-16        2

2026-08-26        2

2026-06-17        6

2026-03-18        9

2026-02-18        10

Agendas & Notes

2026-10-21

Chairs: Benjamin Berk (Care Evolution), Moni Muñoz Torres (Univ. of Colorado Anschutz)

Link to meeting recording:

Actions and meeting summary:

2026-09-16

Chairs: Benjamin Berk (Care Evolution), Moni Muñoz Torres (Univ. of Colorado Anschutz)

Attendees:  Carlos Cardenas-Iniguez, Mousumi Ghosh, Beatrice Amos, Gerald, Edilene Andrade, Jennifer Fostel, Jules Jacobsen

Link to meeting recording: https://us02web.zoom.us/rec/share/PLHX6j2W4hat83F3tC8i3PNzfNkdqi8uy5QHRhIrnbWxJrDmCk3nlU3HSzBsmatB.1-yntBfYmvR-WtYs 

Actions and meeting summary:

Beatrice

• Onboard new attendees (Ravi, Gerald, Edilene) and connect them with relevant groups.

Benjamin

• Send links to the two open PRs (pedigree and kin) in the chat for feedback.
• Review the Phenopacket schema (base.proto, individual.proto, and pedigree.proto) and propose reconciliation for overlapping elements (e.g. individual, age).

Carlos

• Follow up with Moni for suggestions on other journals to submit the Exposome preprint.

Jules

• Send Benjamin the link to the Phenopacket schema (specifically base.proto, individual.proto, and pedigree.proto) for review and reconciliation.
• Provide a couple of sentences or bullet points about Phenopackets updates (e.g. AI readiness) for the Clin/Pheno poster.

Moni

• Organize a conversation between the Phenopackets and Pedigree teams to discuss interoperability, common elements, and removal of redundancies.
• Connect the Exposome team (Carlos and Amy) with the Pan-Canadian Genome Library (Melanie Corto, Jean-Marc Bouquet, Claude Behre) for collaboration.

Quick recap

The meeting focused on updates and discussions around GA4GH standards, particularly the Pedigree, Phenopackets, and Exposome schemas. Benjamin presented updates to the Pedigree model, including the separation of biological and social relationships, the use of external identifiers to reference phenotype and genotype data, and the development of a protobuf representation. Jules and Benjamin discussed potential overlaps and integration between the Pedigree and Phenopackets schemas, considering shared base classes and avoiding redundancy. Moni summarised recent community discussions on Phenopackets, covering topics like survival time, disease severity, and episodic occurrences. Carlos provided an update on the Exposome schema, noting ongoing feedback collection and efforts to engage with global cohorts. The group also discussed preparations for the upcoming GA4GH Connect meeting, including poster and presentation content. New participants Ravi, Gerald, and Edilene introduced themselves, and Beatrice welcomed them to the community.

Connect Meeting Preparation Discussion

The team discussed preparations for an upcoming Connect meeting next week. Moni suggested sharing model updates with the audience, which could be distributed as notes afterward. Benjamin was expected to share updates on Pedigree work.

Pedigree and Kinontology Model Updates

Benjamin presented updates to the pedigree and kinontology models, including the ability to represent relationships in two ways (typical relationships and tree structures), separation of biological and social relationships, and external system linking capabilities. He shared two open pull requests for review and requested feedback from technical, informatics, and clinical perspectives. When asked by Moni about external identifiers, Benjamin clarified that the intention is to reference information in external systems rather than bringing the data directly into the pedigree file.

Protobuf Pedigree Model Integration Discussion

Benjamin presented a new protobuf representation for the pedigree model that includes both conceptual and actual model elements. Jules identified potential conflicts with existing Phenopackets schema elements, particularly around gestational age and individual information. The group discussed creating a common namespace that both schemas could share rather than duplicating elements, with Jules suggesting they could import from each other's schemas while maintaining separate namespaces. Benjamin requested that Jules review the Phenopackets proto and propose reconciliation through comments or commits to address the overlapping elements.

Phenopacket Schema V2 Review

Benjamin was instructed to review Phenopacket Schema V2, focusing on base.proto, individual.proto, and pedigree.proto. Beatrice mentioned connecting Benjamin with Laurie Metz regarding donor conception community involvement and discussed the International Standards Organization's interest in putting pedigree through the product approval process. The team reviewed recent community discussions about Phenopacket improvements, including survival time fields, disease modeling, temporal fields for episodic occurrences, and semantic linking of schema elements, with plans to move forward with implementing changes after final community feedback.

GA4GH Standards Integration Discussion

The team discussed challenges around integrating Phenopackets and Pedigree standards, with Jules highlighting the need to decide whether these should work as a single system or remain separate but interoperable. Moni announced plans to bring together teams to address redundancies and improve interoperability across GA4GH standards. Carlos provided updates on the exposome standard development, noting they are collecting feedback and have scheduled additional meetings to incorporate input from different time zones, including APAC region representatives. The team also discussed preparations for the upcoming GA4GH Connect plenary in Singapore, including poster submissions and welcome session presentations, with several new participants being welcomed to the group.

(Zoom AI generated notes)

2026-08-26

Chairs: Benjamin Berk (Care Evolution), Moni Muñoz Torres (Univ. of Colorado Anschutz),

Attendees: Justina Chung, Peter Robinson, Carlos Cardenas-Iniguez, Coline Thomas, Grant Wood, Jinny Park, Mousumi Ghosh, Orion Buske, Yaqoub Ashhab, Beatrice Amos, Angela Page

Link to the previous latest update: GA4GH Clin/Pheno Data Capture Meeting Minutes 2023-2025

Actions and meeting summary:

1. Prenatal Phenopackets / Beacon (Peter Robinson)

• A group (Peter's team, Teresa Sparks's team, Michael Duzyend's team) has secured curation funding - currently building ~1,000 prenatal phenopackets, targeting completion by March, to serve as a proof-of-concept research cohort.
• Goal: use this as a real-world test case for Beacon + Phenopackets interoperability. Beacon currently lacks a fully functional reference implementation — Peter hopes to collaborate with Michael Baudus (who has a working Beacon-schema database) and Jordi's Beacon team.
• Key technical gap flagged: Beacon API can't handle similarity-based queries (e.g. matching phenopackets by a set of HPO terms with a defined algorithm/threshold) - only simple queries. Peter wants to explore extending Beacon API for this.
• Funding for curation runs out in ~6 months; software development would rely on existing grant FTE time, so scope will stay small.
• Moni offered to connect Peter with the chair of OB at Colorado Children's Hospital, who has budget for AI-driven data-processing work, as a potential collaborator/funding source for prenatal phenopackets.
• Peter also mentioned talks with Jonathan Berg (ClinGen) about ClinGen curating phenopackets, and a plan to build a web version of "Phenoboard" (his phenopacket curation tool, originally built for HPO) - open to collaborators.

2. Phenopackets Community Updates (Moni)

• Ongoing community calls addressing: better representation of survival time, clearer distinction between disease vs. diagnosis fields, expanding temporal fields, and semantic linking of schema elements.
• Encouraged wider community participation - meeting notes and GitHub tickets are available for offline feedback. Next call is in ~2 weeks (around Sept 10).

3. Pedigree Updates (Benjamin Berk)

• Two open PRs ready for review/testing:

1. Pedigree schema - now includes a proto (implementable) version plus auto-generated documentation.
2. KIN ontology v1 - adds new KIN terms, separates biological vs. social relationships, and includes an initial mapping to the HL7 FHIR family relationship ontology.

• Requesting feedback via GitHub comments (or Google Docs, if preferred/no repo access).
• Recommended starting point for reviewers: Read the Docs first, then the OWL/Proto files or CSV (for Kin) depending on familiarity.
• Discussion with Grant Wood on FHIR/Epic interoperability: Epic's FHIR "family member history" only supports individual relationship statements, not full pedigree construction (missing parent-ID linking) - a known historical HL7 gap. Benjamin's model is designed to support both tree-based and named-relationship representations, with an eventual validator planned. Long-term hope: Epic could adopt the GA4GH pedigree model internally with FHIR as a wrapper.
• Plans to add a mapping spreadsheet (Pedigree ↔ FHIR ↔ Phenopackets) to the PR docs, and eventually an implementation guide/translation software.
• Aiming to put Pedigree through the PDAP process to become an official GA4GH standard (timeline TBD).

4. Exposure/Exposome Update (Carlos Cardenas-Iniguez)

• No meetings in July/August; next meeting is next Friday.
• Community asked to implement the draft exposure schema against two exposures (air pollution + one other) using their own cohorts — at least 4 cohorts have expressed interest.
• Next meeting will gather user-experience feedback to shape user stories and refine the draft schema.
• Amy's white paper (foundational rationale for the exposure schema work) was rejected by a couple of venues; still shopping for a publication home, but a preprint exists (Carlos to share the link).

5. PDAP (Product Development & Approval Process) v2.0 Roadshow (Angela Page, with Jinny Park & Justina Chung)

• New process introduces: earlier review checkpoints, risk tiering (based on interoperability risk), a 14-day consent window for Product Steering Committee objections, a formal objection/resolution process (objector must justify late-stage objections; unresolved objections go to a vote), and semantic versioning (only breaking changes trigger a full restart at Stage 0/1).
• Rationale: Steering Committee has historically been under-engaged early on, leading to late objections after years of work; this process forces earlier, more deliberate engagement from the wider community.
• A "product dossier" template (with clear guidance on requirements/questions) is still being finalized and will be shared soon - meant to make the process more concrete than the current high-level slides.
• Peter pushed back hard: found the process overly complex, asked for a 5-sentence summary and concrete examples of past failures to justify the complexity; found terms like "interoperability" too vague without examples.
• Grant Wood offered a defense from HL7 standards experience - argued GA4GH's process is already far more efficient than HL7's, and that this rigor is what gives external (non-member, commercial) organizations confidence to adopt GA4GH standards.
• Jinny Park noted TASC is developing a formal definition of interoperability and accompanying guidance.
• Peter suggested TASC provide concrete good/bad examples for each abstract concept used in the process, to make it actionable.

Next Meeting: September 16th

(AI generated)

2026-06-17

Chairs: Benjamin Berk (Care Evolution), Moni Muñoz Torres (Univ. of Colorado Anschutz),

Attendees: Orion Banks, Grant Wood, Marion Shadbolt.

Link to the latest previous update: GA4GH Clin/Pheno Data Capture Meeting Minutes 2023-2025

Link to meeting recording: https://us02web.zoom.us/rec/share/7maTuykZNMWFR_pFv6N0SzvylBgCn5CECE68kyWFkGF5m0pQgBWzVuluH3isQFEU.Smy1GCBdzbzIn9sO 

Actions and meeting summary:

(AI generated)

• 1. Benjamin: Create and submit a PR for the updated kin ontology and review inconsistencies in the inverse relationships in the OWL file.
• 2. Benjamin: Share the draft documentation (e.g., as a PDF or via read-the-docs draft build) with the group, including Grant, and collect feedback via email/PR comments.
• 3. Moni: Investigate who to contact at EBI to submit the updated KIN ontology to the OLS (Ontology Lookup Service).
• 4. Benjamin: Reconnect with Matt Mullis to ensure he reviews the mapping/tool and provide feedback from the Epic side.
• 5. Moni: Review existing tickets created by Marion regarding schema mapping/conversion challenges, address them, and follow up with Marion as needed.
• 6. Moni: Send out a community engagement email with links to tickets and documentation for asynchronous feedback on PhenoPackets and related infrastructure improvements.
• 7. Benjamin: Consider developing a simple Python script (or similar tool) for converting Pedigree/KIN data to FHIR/HL7 format using the new mappings.
• 8. Moni: Encourage the exposome standard development group to create a repository for tracking tickets, discussions, and schema updates.
• 9. Benjamin: Delete the outdated comment/thread regarding collaboration with R.O. in the repository.

—-----

Key Updates from Benjamin Berk

• Pedigree Documentation & Protobuf Schema: Updated documentation and created a new Protobuf schema (using Proto3, consistent with PhenoPackets/PhenoPac). Used Claude for initial generation; still needs detailed review. PR shared in chat.
• Model Changes (visualised in an updated diagram):

• Split relationship into biological relationship and social relationship (both based on KIN).
• Added egg parent and sperm parent (tree-based references to individuals) - received very positive feedback for simplicity.
• Added twin group and consanguinity.
• Removed karyotypic sex and affected.

• Relationship Direction & Consistency:

• Tree-based (egg/sperm parent) is preferred and should take precedence if inconsistencies arise.
• Standardized on “is-a” relationships only (eliminated “has-a”).
• Added comprehensive ascendant/descendant terms (e.g., grandparent/grandchild).
• Guidance: Prefer relations from the proband (if present); otherwise prefer downward direction and consistency within a pedigree.

• KIN Ontology: Updated in OWL format with inverse properties. A PR is pending. Some inconsistencies noted (e.g. term IDs) that Benjamin will fix during review.
• FHIR/HL7 Mapping: Updated SSSOM mappings. FHIR uses a single relationship field, so biological + social will collapse for compatibility. Easy conversion expected; potential Python script for KIN/Pedigree ↔ FHIR.

Validator & Tooling Needs:

• Update pedigree validator to check consistency between tree-based and relationship-based data.
• Potential “simplifier” tool to detect redundancies/overlaps and suggest canonical representations.
• Orion noted increased complexity (nested structures, multiple representation ways for the same relationships), but viewed it positively overall.

Review & Next Steps

• Benjamin will create two PRs (pedigree + KIN) and share a PDF/draft ReadTheDocs build for feedback.
• Moni and Grant invited to review (Grant prefers email/PR comments).
• Suggestions: Push updated KIN to EBI OLS (Ontology Lookup Service); Moni to investigate.

Broader Phenopackets / Clin/Pheno Updates (Moni)

• Stabilising schema with specific tickets on clinical actions, phenomenon refinement, relationships/linkage.
• Community engagement push: Email “love letter” with organized tickets (technical, docs, bugs, infrastructure). Async contributions welcomed via GitHub.
• Project board (Kanban) in progress to replace Google Docs.
• Ongoing: Pheno Packet store growth, validation tools, FHIR-OMOP conversions. Considering hiring a consultant for FHIR work.
• Exposome standard: Fast progress with strong community interest; user stories, white paper, moving toward GA4GH formal review.

Other Discussion

• Marion refreshed on prior mapping challenges (OMOP splintering, semantic precision vs. source fidelity); existing tickets to be reviewed.
• Positive feedback overall; appreciation for Benjamin’s work.
• Light personal note: Benjamin mentioned recent job loss (company revenue challenges), met with empathy.

Overall Tone: Collaborative, enthusiastic, and forward-moving. Strong praise for the egg/sperm parent modeling and documentation clarity. Focus now shifts to review, validation enhancements, community input, and harmonisation with other standards (FHIR, etc.).

2026-03-18

Chairs: Benjamin Berk (Care Evolution), Moni Muñoz Torres (Univ. of Colorado Anschutz),

Attendees: Beatrice Amos,

Previous meeting minutes: GA4GH Clin/Pheno Data Capture Meeting Minutes 2023-2025

Link to the previous latest update: GA4GH Clin/Pheno Data Capture Meeting Minutes 2023-2025

Link to meeting recording: https://us02web.zoom.us/rec/share/baZoHViVg7PWYPDqZ2ZatLCtaqfKftRSFkMPlPhKdVxkEoxWJzTt7K4ETSf36GGj.XsEhpzwEFiCzL5Kh 

Meeting summary:

Adam presented RareLink, an open-source framework designed to enable interoperability in rare disease research by bridging commonly used but fragmented systems such as REDCap, FHIR, and Phenopackets. Motivated by real-world challenges in capturing and analysing rare disease data, RareLink introduces an ontology-driven common data model that standardises data elements using resources like SNOMED CT, HPO, and LOINC. This model allows data collected in REDCap - widely used but often unstructured - to be transformed into interoperable formats, enabling seamless export to Phenopackets for analysis and integration with FHIR-based clinical systems.

A key contribution of the work is making complex interoperability accessible to clinicians and researchers with limited technical expertise, through tooling, documentation, and automation (including LinkML-based mappings and a command-line interface). RareLink has already been implemented across multiple international sites and use cases, demonstrating its practicality and scalability. Adam also highlighted ongoing challenges, including aligning with formal standards processes (e.g. HL7), improving tooling, and advancing integration with initiatives like the International Patient Summary.

Overall, the presentation emphasised a pragmatic, bottom-up approach to interoperability that prioritises usability and real-world adoption while remaining aligned with broader standardisation efforts.

—

Benjamin presented updates to the GA4GH pedigree model, describing efforts to evolve it from a conceptual framework into a formal standard driven by a real-world implementation with Dana-Farber Cancer Institute. The proposed changes aim to improve clarity, flexibility, and interoperability, including separating biological and social relationships, supporting both tree-based and direct relationship representations, and introducing external identifiers to link pedigree data with systems like FHIR, Phenopackets, and OMOP.

Additional refinements include clearer terminology (e.g. gender identity, sex assigned at birth), handling of complex relationships (e.g. donors, twins, consanguinity), and potential removal of ambiguous fields like “affected” in favour of external phenotype standards. The team is drafting these updates as pull requests, alongside validation tools to ensure consistency across representations, with plans to test through implementation and gather community feedback before progressing through the GA4GH standardisation process.

Broader discussion highlighted challenges with sourcing family data from EHRs, the importance of accommodating diverse family structures and language, and the goal of making the model both computationally robust and practical for real-world adoption.

2026-02-18

Chairs: Benjamin Berk (Care Evolution), Moni Muñoz Torres (Univ. of Colorado Anschutz),

Attendees: Beatrice Amos, Orion Buske, Bradford Powell, Carlos Cardenas-Iniguez, Zoreh Talebizadeh, Omar Khan, Grant Wood, Adriana Malheiro, Jules Jacobsen, Kaitlyn Brethour, Jinny Park, Peter Robinson, Nowrin Aman, Susan Trapp, Gustavo Ybazeta, Nuria Queralt Rosinach, Dominika Bajguz

Previous meeting minutes: GA4GH Clin/Pheno Data Capture Meeting Minutes 2023-2025

Link to the previous latest update: GA4GH Clin/Pheno Data Capture Meeting Minutes 2023-2025

Link to meeting recording: https://us02web.zoom.us/rec/share/3TEmwe65V6ZUEsGkLJuNCcy1O4_7ExdDf4jF-Rsvg44g-c38tqecQe4Sc6pF0ojU.LLAA_HWcNGmIgTN3 

Meeting summary:

Moni’s presentation [link to slides]:

The goal is to explore how existing standards are being used across the global landscape and how they can better support precision medicine, especially given the anticipated growth in genomic sequencing and the need for high-quality, computable phenotype data linked to genotypes.

Moni said the clinical data standards ecosystem is very fragmented.  Important to understand how models such as HL7 FHIR, OMOP, and CDISC complement one another despite differing philosophies. Differences range from clinical workflow exchange to population research and regulatory trial reporting.

Phenopackets is a research-centric, ISO-approved schema optimised for deep phenotyping and genomic diagnostic.

Moni reviewed real-world adoption across institutions, biobanks, and national programs, illustrating varied implementations depending on research and healthcare goals.

Moni also described ETL and data-harmonisation pipelines that transform raw clinical data into standardised, research-ready formats, including REDCap-based frameworks and machine-learning-supported mapping platforms.

She mentioned some of the tools for creating, validating, and converting Phenopackets.

Moni continued by focusing on the growing maturity of interoperability work and the need to strengthen technical and semantic mappings to avoid data silos. She reviewed known cross-standard mapping efforts, including OMOP-to-Phenopackets projects (such as Columbia University’s ontology-supported transformations and COVID-19 Biohackathon exporters), and ongoing FHIR-to-Phenopackets work through HL7’s Vulcan Accelerator, which aims to create implementation guides and semantic mappings. Additional collaborations include joint HL7–CDISC mapping initiatives and multi-standard harmonisation proofs of concept using ontology-driven frameworks. She noted persistent challenges, context-dependent mappings, disease vs. phenotype distinctions, and semantic ambiguities, emphasising the value of collaborative frameworks to address them.

She then outlined a proposed roadmap for the community: consolidating and documenting mappings via a machine-readable registry; convening multi-standards dialogue and potential task forces across SDOs; advancing and benchmarking transformation tooling; piloting native Phenopackets capture at point of care with EHR and research platforms; and expanding scope to areas like pharmacogenomics and longitudinal data, potentially informing future schema versions. She reflected on the broader shift toward decentralised data architectures with shared semantic standards, noting progress in adoption but highlighting the “last mile” challenge of automated, high-fidelity data extraction - requiring stronger ontology use, ETL refinement, and cross-organisational collaboration.

Omar (NHS England) explained that his team’s work is primarily clinically oriented rather than genomics-first, so most of their current APIs are built on the FHIR standard to support clinical workflows. Within NHS England, several digital genomics services are being developed, starting with order management systems, which rely on FHIR but also need to transmit detailed phenotypic data equivalent to Phenopackets for testing and interpretation.

Because of this, he emphasised the value of a consolidated registry of interoperability mappings, which would allow teams like his to reuse existing work rather than build new FHIR-to-Phenopackets mappings themselves. He also noted parallel development of more genomics-specific infrastructure, such as genomic data access and management APIs designed to enable discovery and querying of cloud-hosted genomic datasets, where GA4GH standards like DRS and htsget are being used.

Benjamin: how is FHIR-to-Phenopackets conversion currently handled and are transformation approaches standardised across NHS England?

Omar replied that there is no central guidance yet, individual trusts are choosing their own methods, though there is interest in leveraging existing solutions (e.g. REDCap-based tools) rather than building mappings independently. He noted that current efforts like the HL7 Vulcan Phenopackets Exchange remain limited, mapping only specific elements (such as phenotypic features to Observations) rather than full multi-resource Phenopacket representations, highlighting a need for broader collaboration to advance this work.

Peter added that his team is prototyping FHIR-to-Phenopackets software for upcoming projects and suggested coordinating efforts to create shared, synergistic solutions. He described a longer-term vision involving mining clinical notes via FHIR APIs, combining text extraction with lab and structured data to generate comprehensive Phenopackets for diagnostics and machine-learning use cases.

Moni welcomed this, reinforcing the goal of centralising community work to produce robust, validated tooling. Benjamin echoed interest in narrative-to-FHIR pipelines, noting his own work extracting entities from clinical documents to populate FHIR and downstream models like OMOP.

Zohreh introduced RareX, a survey-based rare disease data platform capturing symptom-focused, patient-reported data across dozens of conditions. She highlighted the need to standardize and map patient-reported outcomes, distinct from EHR or research data, to enable linkage with other registries and avoid fragmentation. Benjamin responded by pointing to the OHDSI OMOP survey working group as an existing forum addressing patient-reported data standardisation [Link].

Clin/Pheno updates:

Pedigree Update:

Benjamin described ongoing work to develop mappings between the Progeny pedigree schema (used at Dana-Farber), Phenopackets, FHIR (as used by Epic for family history data), and the Phenotips schema. This includes enhancements to pedigree and kinship ontology models, with plans to release an updated pedigree version for community feedback.

Exposome Update:

Carlos shared early progress from the Human Exposome data standards effort, building on prior landscape mapping by gathering cohort-level metadata, schemas, and standards from partners to support future ontology harmonisation.

Phenopackets Update:

Moni provided updates from the Phenopackets group, including a recent walkthrough on building Phenopackets step by step and a demo of “Phenoboard,” a cohort curation app, alongside ongoing documentation improvements and feature refinement toward a future version 3.

Moni closed by assigning “homework” to participants: review their data capture schemas, sources, and use cases, identify gaps, and contribute to collaborative efforts around interoperability registries, mappings, and tooling development.