using semantic sql in R: a vignette

[vjcitn]

Users of semantic sql will find a reasonable fraction of the jupyter tutorial
reflected in R/RSQLite operations managed in the ontoProc2 package, which
may appear in Bioc 3.23 in April. For now, github-only:

https://vjcitn.github.io/ontoProc2/articles/ontoProc2.html

[github-actions]

Summary

This issue is a community contribution from @vjcitn (Vince Carey, Bioconductor core), announcing that the ontoProc2 R package provides an R/RSQLite interface to semantic-sql databases. The package is expected to appear in Bioconductor 3.23 (April 2026), with the current version available on GitHub only.

---

What ontoProc2 Does

ontoProc2 is a modernized Bioconductor package that wraps semantic-sql SQLite databases for use in R workflows. It mirrors a significant portion of the semantic-sql Jupyter tutorial but in R idioms.

Key capabilities demonstrated in the vignette:

Feature	Details
DB connection	semsql_connect() — connects to pre-built semantic-sql .db files
Table access	Queries the 100+ views/tables via dplyr + DBI
Label search	Pattern matching against rdfs_label (e.g., find "apoptosis" terms)
Direct edges	Subject–predicate–object graph traversal
Entailed edges	Uses entailed_edge table for inferred (transitive closure) relationships
Cross-ontology	Joins across CL, UBERON, GO, RO, CHEBI

Headline example:

Find all forebrain neurons — filters 185 CL cell types located in forebrain that are subclasses of neuron, returning specialized neuron types across brain regions.

R packages used:

ontoProc2, dplyr, DBI, DT, S7, ontologyIndex

Ontologies demonstrated:

Gene Ontology (GO), Cell Ontology (CL), UBERON, Relational Ontology (RO), ChEBI

---

Significance

This is notable as it substantially lowers the barrier for Bioconductor/R users to work with OBO ontologies via semantic-sql — a large community that previously had no first-class path to the semantic-sql infrastructure. The entailed_edge table (pre-computed transitive closures from relation-graph) is particularly powerful for graph traversal in R without needing Python/OAK.

Worth considering: linking this from the semantic-sql README or tutorial index once the Bioc 3.23 release lands.