Skip to main content

A python implementation of phenopackets protobuf

Project description

Phenopacket schema

Build Status Maven Central Documentation Status

This has been produced as part of the GA4GH Clinical Phenotype Data Capture Workstream and it merges the existing GA4GH metadata-schemas work with a more focused model from the phenopacket-reference-implementation.

This is a re-defined version of the original phenopacket with a more individual-centric approach. This new approach was taken in order to simplify the code required to represent and manipulate the data and also better represent this sort of data as it is in day to day use.


The core documentation can be found at Documentation

The documentation in this README is primarily for the users of the phenopacket-schema java library.

Scope and Purpose

The goal of the phenopacket-schema is to define the phenotypic description of a patient/sample in the context of rare disease or cancer genomic diagnosis. It aims to provide sufficient and shareable information of the data outside of the EHR (Electronic Health Record) with the aim of enabling capturing of sufficient structured data at the point of care by a clinician or clinical geneticist for sharing with other labs or computational analysis of the data in clinical or research environments.

The schema aims to define a common, limited set of data types which may be composed into more specialised types for data sharing between resources using an agreed upon common schema (as defined in base.proto).

This common schema has been used to define the ‘Phenopacket’ which is a catch-all collection of data types, specifically focused on representing rare-disease or cancer samples for both initial data capture and analysis. The phenopacket is designed to be both human and machine-readable, and to inter-operate with the HL7 Fast Healthcare Interoperability Resources Specification (aka FHIR®).


The library uses semantic versioning. See for details.

Email list

There is a low-volume mailing list for announcements about phenopackets at More information about this list is available at


The Phenopacket schema is defined using Protobuf which is “a language-neutral, platform-neutral extensible mechanism for serializing structured data”. There are two ways to use this library, firstly using the Phenopacket as an exchange mechanism, secondly as a schema of basic types on which to build more specialist messages, yet allow for easy interoperability with other resources using the phenopackets schema. The following sections describe how to achieve these two things.

Include phenopackets into your project

Java people can incorporate phenopackets into their code by importing the jar using maven:


Using phenopackets in Python is also straightforward:

pip install phenopackets

Exchanging Phenopackets directly

Examples on how these can be used can be found in the test directory. There are no explicit relationships defined between fields in the phenopacket (apart from the Pedigree), so it is vital that resources exchanging phenopackets agree on what is valid and what the fields mean in relation to other fields in the phenopacket. For example the Phenopacket.genes field may be agreed upon as representing the genes for a gene panel in one context, or a set of candidate genes or perhaps a diagnosed causative gene.

JSON/YAML formats

A Phenopacket can be transformed between the native binary format and JSON using the JsonFormat class from the protobuf-java-util library. This will also need to be added to your pom.xml

pip install protobuf

protobuf-java-util for java and protobuf for python contain simple utility methods to perform these transformations. Usage is shown here:

// Transform a Phenopacket into JSON
Phenopacket original = TestExamples.rareDiseasePhenopacket();

String asJson = JsonFormat.printer().print(original);

// Convert the JSON back to a Phenopacket
Phenopacket.Builder phenoPacketBuilder = Phenopacket.newBuilder();
JsonFormat.parser().merge(jsonString, phenoPacketBuilder);
Phenopacket fromJson =;

// Convert the JSON into YAML (using Jackson)
JsonNode jsonNodeTree = new ObjectMapper().readTree(jsonString);
String yamlPhenopacket = new YAMLMapper().writeValueAsString(jsonNodeTree);

// Convert the YAML back into JSON (using Jackson)
JsonNode jsonNodeTree = new YAMLMapper().readTree(yamlString);
String jsonPhenopacket = new ObjectMapper().writeValueAsString(jsonNodeTree);

// And finally back into a Java object
Phenopacket.Builder phenoPacketBuilder2 = Phenopacket.newBuilder();
JsonFormat.parser().merge(jsonPhenopacket, phenoPacketBuilder2);
Phenopacket fromJson2 =;
from google.protobuf.json_format import Parse, MessageToJson
from google.protobuf.timestamp_pb2 import Timestamp
from phenopackets import Phenopacket, Individual, PhenotypicFeature, OntologyClass

# Parsing phenopackets from json
with open('file.json', 'r') as jsfile:
    phenopacket = Parse(Phenopacket(),

# Writing phenopackets to json
with open('file.json', 'w') as jsfile:
    subject = Individual(id="Zaphod", sex="MALE", date_of_birth=Timestamp(seconds=-123456798))
    phenotypic_features = [PhenotypicFeature(type=OntologyClass(id="HG2G:00001", label="Hoopy")),
                           PhenotypicFeature(type=OntologyClass(id="HG2G:00002", label="Frood"))]

    phenopacket = Phenopacket(id="PPKT:1", subject=subject, phenotypic_features=phenotypic_features)

    json = MessageToJson(phenopacket)

Building new messages from the schema

There is an example of how to do this included in the mme.proto file. Here the Matchmaker Exchange (MME) API has been implemented using the phenopackets schema, defining custom messages as required, but re-using messages from base.proto where applicable. Using the above example, perhaps the Phenopacket.genes is a problem as you wish to record not only the gene panels ordered, but also the candidate genes discovered in two separate fields. In this case, a new bespoke message could be created, using the Gene as a building block.

Git Submodules

This repo uses git submodules to import the VRS protobuf implementation. You may need to use the following command after cloning/update for things to build correctly:

$ git submodule update --init --recursive


The project can be built using the awesome Takari maven wrapper which requires no local maven installation. The only requirement for the build is to have a working java installation and network access.

To do this cd to the project root and run the wrapper scripts:

$ ./mvnw clean install


$ ./mvnw.cmd clean install

Sign artefacts for release

There is a release-sign-artifacts profile for Java which can be triggered with the command

$ ./mvnw clean install -DperformRelease=true

The Python artefacts are released by running


$ bash release-test


$ bash release-prod

Java, Python and C++ artefacts

Building the project will automatically compile Java, Python and C++ artefacts. The Java jar file can be directly used in any Java project. For Python or C++ the build artefacts can be found at




Other languages will need to compile the files in src/main/proto to their desired language. The protobuf developer site has examples on how to do this, e.g GO or C#. Protobuf also supports a host of other languages.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

phenopackets-2.0.2.post1.tar.gz (19.9 kB view hashes)

Uploaded Source

Built Distribution

phenopackets-2.0.2.post1-py3-none-any.whl (27.0 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page