Key points
- Google Antigravity now includes Cheminformatics Skills for drug discovery tasks.
- The agent reduced a workflow iteration from days to one hour in a demo.
- The system uses RDKit, ChEMBL and PubChem connectors for molecular data.
Google DeepMind published a blog post detailing how its Google Antigravity agent now handles cheminformatics tasks, including searching compounds and proposing novel rigidified macrocycles.confirmed The post describes a workflow that accelerates drug discovery by automating data curation and candidate molecule proposal.The agent automates complex chemical data tasks to speed up the early stages of drug design.The system leverages RDKit and curated databases to perform matched molecular pair analysis and propose rigidified macrocycles.
The figures
The blog post states that a single iteration of the drug design workflow typically takes multiple days. Using the new skills, an agent completed a search through thousands of compounds, performed analysis and proposed molecules with synthetic routes in just one hour. The example drug crizotinib has five major rotatable bonds, while the modified lorlatinib has zero effective rotatable bonds.
Behind the figures
The one-hour timeline comes from a technical walkthrough using real-world lung cancer treatments. The agent searched thousands of compounds and performed matched molecular pair analysis. It then proposed novel rigidified macrocycles with synthetic routes. The comparison of rotatable bonds refers to the structural differences between crizotinib, a first-generation inhibitor, and lorlatinib, a third-generation inhibitor.
What changes, what does not
The agent now handles workflow orchestration, database navigation and scientific reasoning. Deterministic chemistry tools execute the molecular calculations. The system unifies data sources such as IUPAC names, Markush structures and structure-data files. The underlying language model manages the process while grounded tools handle the precise chemistry.
What builders can do with it
Developers can build agents that access local tooling for computational and experimental chemists. The skills allow integration with RDKit, an open-source toolkit, and connectors to ChEMBL and PubChem. Builders can create workflows that search compound libraries, perform matched molecular pair analysis and generate synthetic routes. The system supports rigidification strategies to improve drug potency by reducing rotatable bonds.
Why it matters
- The integration of RDKit allows agents to use established, community-vetted cheminformatics tools rather than hallucinating chemical structures.
- Connecting to ChEMBL and PubChem gives agents access to curated, large-scale chemical data without manual scraping or formatting.
- Automating matched molecular pair analysis and synthetic route generation reduces the manual labour required in early-stage drug design.
- The division of labour between the language model and deterministic tools addresses the high error margin in molecular representation.
Topics: models, tools, research topics, vendors
Updated
Then / now / next
Drug discovery workflows required days of manual data curation and model training for each iteration.
Google Antigravity agents can search compounds and propose molecules with synthetic routes in an hour.
Developers may integrate these skills into local tooling for computational and experimental chemists.
Story timeline
- Then Drug discovery workflows required days of manual data curation and model training for each iteration.
- 03 Nov 2026 Agentic cheminformatics with Google Antigravity
- Now Google Antigravity agents can search compounds and propose molecules with synthetic routes in an hour.
- Next Developers may integrate these skills into local tooling for computational and experimental chemists.
The strongest doubt
Counterpoint from the sources
The blog post notes that altering a single atom or bond describes a different compound, highlighting the zero margin for error in these workflows.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Questions readers ask
What tools does Google Antigravity use for cheminformatics?
It uses RDKit for calculations and connects to ChEMBL and PubChem for data access.
How does the agent handle different molecular representations?
It unifies sources like IUPAC names, Markush structures and SDF files through deterministic chemistry tools.
What is the time saving demonstrated in the blog post?
The agent completed a workflow in one hour that typically takes multiple days for a single iteration.
Connected stories
Reading receipt
- Google Antigravity now includes Cheminformatics Skills for drug discovery tasks.
- The agent reduced a workflow iteration from days to one hour in a demo.
- The system uses RDKit, ChEMBL and PubChem connectors for molecular data.