Upload your differentially expressed genes (DEGs) as a results table from
DESeq2, edgeR, limma, Seurat or GEO2R and pick a disease. Every gene is checked against Open Targets, FDA drug labels,
ClinicalTrials.gov and CIViC, and you get three separate answers: the genes worth an
experiment, the approved drugs that could be repurposed, and what is already being targeted
in that disease — each row linked to the evidence behind it.
Direction-aware
A drug that inhibits a gene you found down-regulated points the wrong way. It is rejected, not ranked.
Not what is already given
Drugs already labelled or in trials for your disease are not offered back to you as repurposing ideas.
Nothing hidden
Rejected rows stay on the page with their reason. Your file is read in memory and never stored; no gene names are logged.
1 · file2 · check the columns3 · report
Drop your DEG table here, or click to chooseCSV, TSV or Excel (.xlsx), up to 8 MB · it is read in memory and never stored
No file at hand? Try a real example (neuroblastoma, 396 genes):
Check how we read your file
Detected from the header row. Correct anything that is wrong — a wrong column
here changes every number below it.
Direction decides which drugs are kept: an inhibitor is only useful against something
that is up. If this is the wrong way round the whole report inverts, so there
is no default.
Methods, data and limits
How the three answers are made
They are kept apart on purpose rather than merged into one number. Genes worth an experiment:
tractable in principle, nothing approved yet; ranked by disease association (45%), your DEG
(differentially expressed gene) statistics (30%) and hard evidence (25%), where hard evidence means genetic association or
somatic mutation only. Known-drug evidence is left out of the score because it is circular in a
score about finding drugs. Repurposing: approved drugs whose target is in your list, after
the direction check and after removing drugs already labelled or in trials for this disease.
Already targeted here: a property of the disease, not of your comparison.
Direction is a gate, not a score
The mechanism of each drug (inhibitor, agonist and so on) is compared with the direction of
your gene. A mismatch removes the drug from the ranking and keeps it visible, with the reason,
as a rejected row.
Data
Open Targets Platform: disease–target association, tractability and known drugs (association tables cached up to 7 days).
openFDA drug labels: which drugs carry a label naming the disease (refreshed weekly).
ClinicalTrials.gov API v2: registered interventional trials; the retrieval date is printed on each disease briefing.
CIViC: variant-level therapy evidence, nightly export (CC0), snapshot of 2026-10-06.
53 disease briefings supply the approved and in-trial context. A disease without one is scored without it, and the report says so.
Limits
Tractable is not effective. A row is a lead for an experiment, not a treatment recommendation.
An FDA label naming a disease is counted, which is not the same as standard of care: drugs approved under broader indications are not visible to a label search.
Association scores reflect published evidence, which is richer for well-studied genes.
Privacy and reuse
Your file is read in memory and never stored. No gene name is written to a log.
To cite: BioTransfer Gene Druggability Analysis, usebiotransfer.org/tools/deg-actionability/, accessed on the date you ran it.