BindingDB

Description: BindingDB is a public, web-accessible database of measured binding affinities, focusing chiefly on the interactions of protein considered to be drug-targets with small, drug-like molecules. As of July 31, 2021, BindingDB contains 41,300 Entries, each with a DOI, containing 2,303,972 binding data for 8,561 protein targets and 995,797 small molecules.

Cross-references: https://github.com/NCATSTranslator/Translator-All/wiki/bindingdb

Synonyms: The Binding Database

Edge Categories Distribution

Epistemic Robustness

Epistemic Robustness measures the provenance quality of edges from this knowledge source, using Knowledge Level and Agent Type from the Biolink Model to assess evidence strength and reliability. Higher scores indicate stronger evidence and greater manual curation.

1.00
Average Epistemic Score

1.00
Knowledge Level Score
Most common: Knowledge Assertion

1.00
Agent Type Score
Most common: Manual Agent

1,636,404
Edges Included

0
Edges with Missing Provenance
(Both Knowledge Level and Agent Type not provided)

Edge Type Validation

This section shows the validation of edge types in this knowledge source against the recognized biolink schema patterns.

1,636,097
Recognized Edges
( 100.0 %)

0
Unrecognized Edges
(0.0%)

ABox / TBox Classification

The TBox-to-ABox balance reflects how much a knowledge graph emphasizes abstract schema versus concrete instances—too much of either can hinder effective learning and reasoning.

1,245,964
ABox Edges
( 76.2 %)

0
TBox Edges
( 0.0 %)

390,133
Undefined
( 23.8 %)