Labs
Hands-on tools built on the logic of open-source science projects. Click a building, or pick a card. Everything runs in your browser; AI extras use your own key.
Isotope vans really do carry Lu-177 and F-18 around this map. Pixel art drawn in code, in the spirit of OpenTTD.
Open-source foundations
- radioactivedecay (opens in a new tab) (MIT): Decay chains, branching and ICRP-107 data used by the decay engine
- ISOTOPIA (opens in a new tab) (see repository): Medical isotope production yield (activation equation)
- IsotopeRelay (opens in a new tab) (MIT): Half-life-aware delivery windows for shipments
- RadSheet (opens in a new tab) (proprietary (concept only)): Decay manifests keyed to departure and arrival
- PyRIID (opens in a new tab) (BSD-3-Clause): Gamma-spectrum identification and proportion estimation
- RadioGx (bhklab) (opens in a new tab) (see repository): Linear-quadratic radiation response, SF2, AUC, D10
- PharmacoGx (opens in a new tab) (GPL-3.0): Log-logistic dose-response fitting, AAC and IC50
- PharmacoDB (opens in a new tab) (GPL-3.0): Live cell-line dose-response data (GraphQL API)
- PubChemPy (opens in a new tab) (MIT): PubChem PUG REST lookups
- REINVENT4 (opens in a new tab) (Apache-2.0): Desirability scoring and generative design loop
- DeepChem (opens in a new tab) (MIT): Molecular featurization ideas behind the in-browser descriptors
- TorchDrug (opens in a new tab) (Apache-2.0): Graph view of molecules; property-prediction framing
- GenAIDD (opens in a new tab) (see repository): Generative models in drug design
- GitHub topic: ai-drug-design (opens in a new tab) (various): Survey of AI drug design projects
- PCGR (opens in a new tab) (MIT): Variant tiering (AMP/ASCO/CAP) logic
- openoncology (opens in a new tab) (MIT): Ranking therapy candidates from variant evidence
- OnCo (opens in a new tab) (MIT): Cited oncology knowledge graph framing
- OpenTTD (opens in a new tab) (GPL-2.0): Visual and gameplay inspiration (no OpenTTD assets are used)
- open-source-games (opens in a new tab) (list): Inspiration for the pixel and game style
RadioGx re-implements ideas and published models in JavaScript; it does not bundle these projects’ code, except the radioactivedecay dataset (ICRP-107 / AME2020 / NUBASE2020, distributed under its licence). Labs are for research and education, not clinical dosimetry, transport compliance or treatment decisions.
BYOK AI research modules
Connect your provider once in AI settings, then move from evidence to inference, synthesis and a testable research plan.
About this tool
What is Labs?
RadioGx Labs are interactive teaching calculators for radiopharmacy, radiobiology, pharmacology and genomics: decay chains, production yield, shipment windows, gamma-spectrum identification, the LQ model, dose response, molecules and variant tiering, plus the Isotope Logistics game.
Key features
- Eight labs built on named open-source methods
- Real ICRP-107 decay data through radioactivedecay
- Teaching values, never patient dosimetry
Frequently asked questions
Can I use the Labs for patient dosimetry?
No. The labs are teaching calculators. Patient-specific dosimetry and radiation safety decisions need validated clinical tools and qualified medical physicists.