Automat

Back to Home

Overview

Automat was created by the Ranking Agent team. Automat hosts a diverse set of knowledge sources, represented as knowledge graphs (KGs) and derived largely from curated data sources.

Each knowledge source ingested into Automat is:

• Normalized using Node Normalization and Edge Normalization.
• Loaded into its own independent Neo4j database, so that it can be independently updated and maintained.
• Exposed using the SRI's Plater tool, which stands up the following interfaces automatically given the Neo4j database.

Knowledge Sources

Descriptions of each Automat KG are provided below by way of links to GitHub webpages. Each description includes:

• a high-level overview of the knowledge source, including the types of edge assertions that are supported;
• a version number;
• a URL;
• a Neo4j endpoint; and
• counts of nodes and edges in the KG.

---

• ROBOKOP KG
• Alliance of Genome Resources
• CAM-KP
• CTD
• DrugCentral
• FoodDB - Not yet available, see https://foodb.ca/ for additional information
• GTEx
• GtoPdb
• GWAS Catalog
• Hetio
• HGNC Gene Families
• HMDB
• Human GOA
• ICEES
• IntAct
• PANTHER
• Pharos
• RENCI SRI Reference KG
• STRING DB
• SRI Reference KG
• Text Mining Provider KG
• Ubergraph *Also see , https://github.com/NCATSTranslator/Translator-All/wiki/SRI-Ontology-Service
• Viral Proteome

Interfaces

• Reasoner API v1.4.0 query
• Reasoner API v1.4.0 overlay
• SmartAPI-style one-hop queries
• Neo4j Cypher via a HTTP POST

Source Code

• https://github.com/TranslatorSRI/Plater
• https://github.com/ranking-agent/reasoner
• https://github.com/RENCI-AUTOMAT/

External Documentation