GEOmend

GEOmend combines metadata from Gene Expression Omnibus (GEO) studies and local tables into one table for analysis. It suggests which columns describe the same variable, converts compatible units, and keeps the sources and decisions behind the result. You can export a first table, then correct the matches that matter to your analysis.

Produce a first table

With Python 3.13 or newer:

python -m pip install geomend
geomend capabilities --json
geomend project create samples-a.tsv samples-b.tsv --project my-project
geomend project exports table my-project --mode best_guess --profile analysis --output metadata.tsv

Replace the filenames with CSV/TSV files, local family-SOFT files, or GEO identifiers such as GSE5390. GEO identifiers require an initial metadata download. The export writes metadata.tsv, column descriptions in metadata.columns.tsv, and extra metadata in metadata.context.tsv. It keeps one row per study and sample.

best_guess can include unreviewed suggestions. Inspect the source evidence for matches that affect your analysis. Choose --mode reviewed to use only reviewed mappings; project tables still keep their sample rows when reviewed values are blank. If a run stops, resume with geomend project inputs run my-project.

Find the instructions for your task

The documentation home explains which workflow fits your goal. geomend project --help shows command groups; --expanded-help shows all options. Record decisions through the CLI, SDK, or authenticated API. Preserve source metadata and review records; direct database edits are unsupported.

To build a GUI, notebook workflow, or agent on GEOmend, use the client integration guide, project operation reference, and reviewer client contract.