bindsight

Expression → Binder. An open-source pipeline that takes RNA-seq counts and outputs ranked de novo protein-binder candidates against over-expressed cell-surface antigens, with full PROV-O / RO-Crate provenance back to the patient cohort.

Start here

• What is bindsight? — the 5-minute pitch.
• How to use it — install, the bindsight demo, and the full discover → design → validate → rank → report → export flow.
• Use cases — concrete scenarios.
• Designing on Colab — the GPU half on free Colab.

One-command demo

pip install -e ".[discover,report]"
bindsight demo

Runs the discovery half on a real TCGA-BRCA cohort (auto-downloaded from NIH/GDC), discovering antibody-tractable surface antigens with full provenance.

How it fits together

The CLI (bindsight …) and an optional Snakemake front-end both drive the same Python pipeline. The discovery half is CPU-only; the design half (RFdiffusion → ProteinMPNN → Boltz-2, plus BindCraft / BoltzGen / Chai-1r / AF2-IG) runs on a GPU backend you choose (Modal / local Docker / Kaggle / Colab).

See ARCHITECTURE.md, LICENSING.md, and CONTRIBUTING.md in the repository for design rationale, the per-component license inventory, and how to add a designer / validator / runner plugin.