Context is the first layer of intelligence.
Reomics begins with auditable metadata curation, then uses that foundation to make public RNA-seq cohorts discoverable and prioritize deeper analysis.
A path from archive to answer.
The platform roadmap is deliberately staged: first establish reliable labels, then make them queryable, then process the cohorts with the greatest research value.
Public sources
SRA · BioSample · GEO · publications
Messy metadata
Unstructured fields and missing context
Harmonization
Extraction · mapping · validation
Cohort tables
Traceable labels and confidence
Evidence
Search, prioritize, investigate
Build the foundation, then build on it.
The first milestone is a validated table of human samples, subjects, tissues, diseases, and experiment types with explicit uncertainty. Subsequent stages build on that shared context.
Metadata curation
Parse source records, extract candidate fields using rules and models, map terms to ontologies, and review samples against a gold set.
Query layer
Make curated tables searchable through structured and natural-language workflows for cohort discovery.
Prioritized processing
Use Monorail pilots to assess high-value disease cohorts and expression or splicing questions before scaling.
Partner-specific evidence
Develop evidence packages and, where appropriate, expand to protected data within approved environments.
Every label needs a trail back to its source.
Reomics intends to store field-level provenance, confidence, extraction rationale, and review status. Models can propose candidate labels; the validated table is the durable resource.
Manual review and public correction workflows are part of the proposed quality loop.
Explore a research question with us.
We welcome conversations with biotechnology and pharmaceutical teams, researchers, and prospective collaborators.