Use cases and competency questions
Competency questions (CQs) are the questions the knowledge graph must be able to answer. Each has a SPARQL query, toy data designed to exercise it, and an expected result checked in CI.
| Use case |
Queries |
Toy data |
| UC1 Aeromicrobiome functional characterisation |
queries/uc1-aeromicrobiome/ |
examples/valid/Container-uc1-aeromicrobiome.yaml |
| UC2 Antimicrobial resistance transmission |
queries/uc2-amr-transmission/ |
examples/valid/Container-uc2-amr.yaml |
| UC3 Host–microbiome interactions in cancer |
queries/uc3-host-microbiome-cancer/ |
examples/valid/Container-uc3-cancer.yaml |
| UC4 Metabolic impact of nutrition |
queries/uc4-nutrition-metabolism/ |
examples/valid/Container-uc4-nutrition.yaml |
UC3 is the most complete one; copy its structure.
UC1: Aeromicrobiome
| ID |
Question |
Tests the model on… |
| uc1-cq01 |
Which GO processes are detected in air samples, per local setting? |
ENVO sample context, gene function, subclass closure |
| uc1-cq02 |
Which air taxa carry stress-survival genes (sporulation, UV, desiccation)? |
gene → taxon, GO hierarchy |
| uc1-cq03 |
Which air taxa have humans as a host? |
bridging observations to taxon–host interactions |
UC2: AMR transmission
| ID |
Question |
Tests the model on… |
| uc2-cq01 |
Which determinants found in waste water appear later in human stool from the same place? |
place and time joins, host taxon via has_host/in_taxon |
| uc2-cq02 |
Which determinants on mobile elements are carried by more than one taxon? |
SO genomic context, aggregation |
| uc2-cq03 |
Which drug classes are affected by determinants in all three One Health compartments? |
compartment enum, ARO drug classes |
UC3: Host–microbiome in cancer
| ID |
Question |
Tests the model on… |
| uc3-cq01 |
Which taxa are associated with colorectal cancer (subtypes and DOID/NCIT equivalents included)? |
disease hierarchy, SSSOM expansion, evidence |
| uc3-cq02 |
Are findings consistent, conflicting, or from a single study? |
reified associations keeping contradictory claims |
| uc3-cq03 |
Which genotoxic or carcinogenic metabolites are produced by CRC-associated taxa? |
chaining association types, ChEBI roles |
UC4: Nutrition
| ID |
Question |
Tests the model on… |
| uc4-cq01 |
Which diets increase butyrate-producing taxa? |
diet → taxon → metabolite |
| uc4-cq02 |
Food → taxon → metabolite → disease: which outcomes, in which direction? |
3-hop chains |
| uc4-cq03 |
Which diet-induced metabolite changes have no known producer? |
gap analysis with FILTER NOT EXISTS |
Proposing a new competency question
A good question:
- names the entities involved (taxon, habitat, disease…) and the context that matters (body site, time, place);
- includes a concrete example answer from your domain, so toy data can be written;
- says which existing resource could supply real data (e.g. MGnify, CARD, gutMDisorder).
Then add the query, toy data and expected result (see Onboarding).