Edge Normalizer

Back to Home

Description

The Biolink Model defines allowed predicates in the Translator ecosystem. Ingesting data from arbitrary sources requires mapping predicates in those sources to Translator predicates.

The Biolink Lookup Service can find predicates if they have an exact mapping in the model. The EdgeNormalization service takes this a step further, and attempts to find the best match to a Biolink predicate, even if there is not an explicit mapping. Usually, this tool will not be called during the operation of a Translator query, but will be used by KP teams during the ingest of data into a Translator resource.

Most users will use the public service, but instructions for deploying a new instance are given below.

Usage

The Edge Normalizer has a single non-TRAPI endpoint, /resolve_predicate, which takes a CURIE representing a relationship, and returns the best match in the biolink model for that query.

CURL Example:

curl -X GET "https://edgenormalization-sri.renci.org/resolve_predicate?predicate=SEMMEDDB%3ACAUSES&version=latest" -H  "accept: application/json"

Response:

{
  "SEMMEDDB:CAUSES": {
    "identifier": "biolink:causes",
    "label": "causes"
  }
}

More examples of calling the service, including calling with multiple input CURIES, are given in the examples notebook.

Deployment

Deployment of the EdgeNormalizer is via docker. Information about building the docker images, and deployment via kubernetes can be found in the github repository

Issues

For reporting or reviewing issues related to Node Normalization, visit the Github Issues.

Links

• Live Interface https://edgenormalization-sri.renci.org/docs
• Github Repository https://github.com/TranslatorSRI/EdgeNormalization
• SmartAPI Registry http://smart-api.info/registry?q=528c38e2218202a1095cc70c612af625