What is tellic data?

Overview of tellic data

tellic data represents the unique data asset generated by our data curation pipeline - knowledgen. This output is a composite graph-relational dataset that indexes key meta-data from structured and unstructured sources. The dataset is architected into what we call our "Hybrid Data Index". This hybrid index is used to empower graph-related analyses while supporting rapid front-end application display and drill-downs.

The Hybrid Data Index consists of two core concepts: nodes and relationships. Nodes are key biological entities that form the vertices in the graph. Relationships form the edges of the graph.

Relationships are derived from three methods:

  • tellic’s advanced transformer deep-learning methods applied to unstructured text
  • Weight values provided in structured data(e.g. GWAS–logP)
  • Weightless values provided from structured data (e.g.indication-intervention relationships from Clinical Trials)
The details of the relationships can be found in the Documentation. In addition, the Hybrid Data Index contains a document-level keyword index. For each document in curated data sets, all detected concepts are tagged along with their frequency of occurrence and counts. This data correlates the node concept ID’s with the document IDs to empower rapid search functionality