​

The Open Targets Platform integrates multiomics and chemical information from public available data sources.

You can inject your own custom or private data following the summary below. The pre-requisites are as follows:

• data contains a target identified by Ensembl gene ID

• data contains a disease or phenotype identified by an ontology term e.g. EFO, HP, Orphanet

• data should conform to our JSON schema

Transform your data into an array of JSON objects

You need to transform your data in an array of JSON objects. The JSON database schema is publicly available on GitHub.

Each object requires mapping to an Ensembl gene ID and an EFO ID (or HP terms, in case the EFO term is missing). We have a light wrapper called OnToma to facilitate this mapping process.

Once the conversion is done, you will have an array containing a JSON object. For instance, our PheWAS evidence will look something like this:

[{"target": {..  "http://identifiers.org/ensembl/ENSG00000130204", }, "sourceID": "phewas_catalog", 
    "variant": {"type": "snp single", "id": "http://identifiers.org/dbsnp/rs2075650"}, 
    "disease": {"id": "EFO:0000249"}, 
    "unique_association_fields": {..}, 
    "evidence": {... , "value": "5.237e-28"}, 
     "type": "genetic_association", "resource_score": {...}}
{"target": { ... } disease: {...} ... }
{"target": { ...}, ...}
...
]

Validate the data

Before processing your data, you should run Open Targets validator to check for any inconsistencies in your data.

After installation, the validator can be run as it follows:

cat file.json | opentargets_validator --schema https://raw.githubusercontent.com/opentargets/json_schema/master/src/literature_curated.json

Run the Open Targets Platform data pipeline (mrTarget)

The Open Targets Platform pipeline parses and integrates multiple files containing evidence, discrete pieces of data linking targets to diseases. It then aggregates them into associations and ranks each target-disease connections according to the quality and quantity of each evidence. To inject your own data (i.e. a set of evidence) to the Open Targets Platform, you need to add it to the pipeline processing.

Briefly - once you have access to the pipeline - you ought to:

• Fork it from GitHub

• Add the additional evidence to the list of Open Targets data sources

• Build a new docker container

• Start the evidence processing with mrtarget --evs