Disease, treated as an equation.
The lab agent never stops. It takes a disease, reads what the world knows about it right now, and scores every existing drug — and every pair of drugs — on how well its mechanism fits the disease. Each score is an equation you can open and read term by term.
Read this first. Every score is a research hypothesis built from public evidence. A high score is not evidence that a drug works, is safe, or cures anything. Nothing here is medical advice, and no candidate has been tested in a laboratory, an animal or a person by this lab.
Run log
Loading- Loading the latest cycle…
Evidence found
Newest first- Loading…
Candidate leaderboard
Scores are 0–100 and change as evidence changes. Click a row to see the equation behind it.
The record column is what has actually happened to the drug in people with this disease, read from ClinicalTrials.gov: the highest phase its trials reached, with a drop when one was stopped for futility, lack of efficacy or a safety signal. The percentage beside it is the industry base rate for that phase — of all drugs that reached it, how many were eventually approved for anything. It is history, not a prediction, and nothing here is simulated.
| # | Candidate | Score | Change | Status | Record | Why |
|---|---|---|---|---|---|---|
| Loading… | ||||||
Disease network
Filled nodes are genes Open Targets associates with the disease; node size follows the association strength. Lines are high-confidence STRING associations. The lab measures distances on this graph to tell complementary drug pairs from redundant ones.
Shared pathways
Reactome- Loading…
How a score is built
Single drug: 26% strongest target association, 22% clinical record in people with this disease, 12% interaction strength, 12% literature, 12% approval status, 8% share of the disease module covered, 8% safety from FDA adverse-event reports. Withdrawn drugs are cut to a quarter and ranked last.
Pair: 40% how strong both partners are alone, 20% coverage the pair adds over the better single, 15% network separation of their targets (the complementary-exposure pattern from network pharmacology), 10% pathway complementarity, 15% evidence they have already been tested together.