Skip to main content

SMA Research Platform

Evidence graph for Spinal Muscular Atrophy

Biology-first target discovery
Christian Fischer / Bryzant Labs
Targets
Trials
Drugs
Datasets
Sources
Claims
Evidence
Hypotheses

Wet-Lab Candidate Shortlist

COMPUTATIONAL HYPOTHESES

This is the ranked, evidence-backed shortlist of potential wet-lab candidatesassembled by the pipeline’s candidate-ranking agent. It does not rank by predicted potency or affinity — it ranks by convergent computational evidence: how well independent in-silico signals (folded structure, docking pose, MD pose-residence, buyability / freedom-to-operate / toxicophore flags) agree for a given target. A high score means “several independent methods point the same way,” which makes it a good starting point for wet-lab triage — not a measured result and not a claim that the molecule works.

The shortlist is organised into three lanes: Compound (small-molecule leads), Mechanism(AF3 multi-chain protein–protein complexes to test biochemically), and De-novo Binder — de-novo designed mini-proteins. The table reflects an upstream shortlist and is not a complete research inventory. Assess each design using its exact identity, available artifacts and run-specific evidence; inclusion here does not establish experimental readiness.

How does candidate ranking work?

Rows come from the candidate_shortlist table, rebuilt continuously by the CQO candidate-ranking agent. Each row is scored on convergent evidence and labelled with an honest confidence tier — promising, moderate, or exploratory — plus the raw evidence breakdown and provenance so you can see exactly how each hypothesis was derived.

Compound lane. Evidence chips include MD pose-residence (fraction of the trajectory the ligand stays in the pocket), hERG risk, toxicophore and freedom-to-operate flags, CNS penetrance, and a catalogue price band. Buyability is called out per row — in-stock, make-on-demand, or custom synthesis.

Mechanism lane. AF3 multi-chain interfaces, chipped with the interface iPTM. These are structurally-supported complexes proposed as mechanistic hypotheses to confirm by co-IP / proximity ligation + SPR.

None of this is experimental validation. candidate_score ranks computational agreement within the pipeline; it is not an IC50, a Kd, or an efficacy claim. Treat every entry as a hypothesis to test at the bench.

Loading candidate shortlist…

Login → Command Center