Science is entering its Agentic Era Introducing the Scientific AI Ecosystem Map! 🗺️ We mapped the tools, models & agentic frameworks powering the new substrate for bio Built to help founders, researchers & builders navigate the explosion in scientific AI Let’s break it down🧵

Scientific AI agents are no longer sci-fi 👾 They can: → Aggregate SOTA research → Formulate scientific hypotheses → Design and execute experiments But every lab is building in silos. That’s why we see a clear need to add to what the web2 Agentic Science world is creating.

To that end, the era of 'collective scientific intelligence' is arriving. From literature mining to lab automation, we've mapped the 5 key layers powering scientific AI and advancing the biological design space👇

1. Knowledge & Data Sources Everything starts with data. Scientific agents pull from: • PubMed, Dataverse (papers & academic data) • PDB, UniProt (proteins) • ChEMBL, PubChem (chemistry) • Reactome, Gene Ontology (pathways) • AACT (clinical trials) • @biorxivpreprint (preprints)

2. Agentic Orchestration Layers These tools bring reasoning to life. Examples: • @ElizaEcoFund - plug-and-play agent framework • ChemCrow - chemistry assistant • BioMCP / PubMedMCP - literature servers • “Coscientist” - autonomous lab agent • @PrimeIntellect - modular agent infra

3. Foundation & Domain Models The high-octane workhorses. Examples: • AlphaFold 2 - structure prediction • Evo 2 - genomic modeling (Arc Institute) • Geneformer - gene expression • ChemGPT - chemistry reasoning • ESM-3, PaLM-2, Claude 3 - language scaffolds

4. Experiment & Simulation Execution AI running experiments. Systems include: • Opentrons - API-driven lab robots • INDRA - mechanistic modeling • PyMOL / AlphaFold inference • Cloud-based simulation: OpenMM, Foldit • SMART-on-FHIR: early EHR integration

5. Scientist-Facing Platforms & Assistants Where AI meets researchers. • @GoogleAI Co-Scientist - hypothesis brainstorming • @FutureHouseSF - Q&A, systematic reviews, chemistry planning • Consensus & Elicit - structured paper reviews • Biomni, Iris, Scite - specialized paper agents • @vita_dao AUBRAI, ChatPaper - AI research co-pilots

Together, these tools form a new research stack: → Models interpret biology → Agents orchestrate reasoning → Platforms deliver assistance → Robotics and simulation close the loop

The Scientific AI Ecosystem Map isn’t just a list of tools. It’s a blueprint to bio/acc 🧪 From search → to synthesis From analysis → to action From static knowledge → to systems that learn