Evaluation for genomickb

Evaluator: Cortes et al.

Evaluated on: 2025-08-26

This is a manual evaluation intended to identify potential barriers to reuse.

Evaluation Criteria: This evaluation uses the KG-Registry evaluation rubric as described in Cortes et al. (2025) . The rubric assesses knowledge graphs across multiple dimensions including access, provenance, documentation, maintenance, and fitness for purpose.


Access Level and Types

QuestionAnswerComment
Access to data outside of the knowledge graphYCan submit an online query and download subgraphs that meet criteria
API or online access to the knowledge graphYWebsite with UI for creating queries, API supposely coming in v2, Neo4j dump available but not hosted by creators
Multiple access options availableYCan use online query tool or download Neo4j dump
Source code availabilityYHttps://github.com/tinalee-tech/KG-dataloader
Downloadable knowledge graphYNeo4j dump

Section Score: 5/5

Provenance of Nodes and Edges

QuestionAnswerComment
Source list providedYSupplementary note 1
Source versions informationN
Import dependenciesN
Node and edge sourcesYTables 1 (nodes) & 2 (edges), query allows users to restrict by datasource
Edges deduplicationYConsidered but not de-duplicated. Users can restric querys to specific datasources and duplicate edges can be used for validation. EX: ‘enhancer regulate gene’ with restriction ‘cell_line=GM12878’ and ‘data_source=EnhancerAtlas’ -> 118,610 node pairs. A second query for ‘variant overlap enhancer’, ‘enhancer regulate gene’, and ‘variant correlate_with gene’ -> 16,871 enhancer-gene pairs from EnhancerAtlas can be validated by GTEx eQTLs
Triples source detailsYTable 2
Edge type schemaYTable 2

Section Score: 5/7

Documented standards, schema, construction

QuestionAnswerComment
Biological usable dataYThey have a web interface (https://gkb.dcmb.med.umich.edu/) presenting the entities and edges in a human-readable format. It can be used to answer genomics-related questions and conduct biological analysis
Resolvable IDsYUses one external ID when possible and defines globaly unique IDs for entities without such as ChIP-seq peaks
Construction documentationYGithub and supplement
Transformation documentationYSup Note 1 Data Collection and Processing
Schema usedYUses custom schema not an existing one but does provide documentation

Section Score: 5/5

Update frequency and versioning

QuestionAnswerComment
Stable versionsYBut still on version 1.0 despite multiple upates listed
Public tracker informationNFAQ says to email the team
Knowledge graph contact informationYWebsite has a contacts button with name/emal for both technical details and accessing the data/web server
Updated annuallyYDuring 2022-2023 yes but not since then
Prior versions accessYClearly documented chages with dates but prior versions not available. *Paper states that they store all versions internally and can revert if unexpected changes are ever made to their graph but these are not accessible publicaly

Section Score: 4/5

Evaluation - Metrics and Fitness for Purpose

QuestionAnswerComment
Use case providedNThe paper provides examples of specific queries that can be performed through the website that would normally require coding and integrating multiple data sources, but it is limited to knowledge extraction and does not show how this could be used to make novel biological discoveries
Evaluation against other modelsN
Defined scopeYCoding free queries of human genomic information including both functional, strutural, epigenetic and other relationships
Multiple evaluation methodsN
Accuracy metricsYKind of but not explicitly. Since this tool is all about data querying, confidence could be assesed by creating multiple queries like the example of using eQTLs to "validate" EnhancerAtlas relationships

Section Score: 2/5

License Information

QuestionAnswerComment
LicenseCC BY-NC 4.0