Google DeepMind has unveiled a suite of Cheminformatics Skills for Google Antigravity, positioning the agent-powered platform as a collaborative science workbench for chemistry. According to the company, workflows that typically take multiple days for a single drug discovery iteration — curating complex data from charts, figures and literature, training predictive models, and proposing candidate molecules to synthesize — can now be completed in about an hour.

In a blog post published on November 3, 2026, researchers Zoi Tsangalidou, Ruijie (Kelly) Zhang, Sukhdeep Singh and Michael V. LeVine described the system using real-world lung cancer treatments as a demonstration: an agent equipped with the new skills searched through thousands of compounds, performed matched molecular pair analysis, and proposed novel rigidified macrocycles complete with synthetic routes in just one hour.

Why Cheminformatics Data Remains Hard to Handle

Chemistry is one of science's richest public records, spanning millions of structures across journal articles and vendor catalogs. It provides a common representation that links physics and mathematics to the life and environmental sciences, and molecules and their transformations are ubiquitous, shaping the everyday world.

Despite decades of progress in cheminformatics, chemists still spend too much time wrangling data. The process is complicated by the fact that molecules can be represented in many ways, from text descriptions to images or data structures: the same compound may appear as an IUPAC name in a paper, a Markush structure in a patent, or a structure-data file (SDF) in a catalog. Advanced cheminformatics systems must be able to unify these sources and deal with the many ways errors can arise in them.

According to the team, this task has no margin for error — alter a single atom or bond and you have described a different compound altogether, rendering a promising inhibitor or catalyst inert.

A Division of Labor Between Language Models and Deterministic Tools

To let scientists focus on hypotheses rather than data wrangling, DeepMind built the Cheminformatics Skills to turn Antigravity's surface into a collaborative science workbench with full access to local tooling for both computational chemists and experimental bench chemists.

The skills also create a division of labor in which the underlying language model handles workflow orchestration, database navigation and scientific reasoning, while deterministic chemistry tools execute molecular calculations. The system is grounded in RDKit, the foundational open source cheminformatics toolkit developed over two decades by a global community of contributors, alongside connectors to curated resources from ChEMBL and PubChem. Together, these components encode domain best practices directly into the agent's workflow.

Rigidifying a Floppy Molecule

The technical walkthrough highlights a classic medicinal chemistry technique for improving drug potency: rigidification.

Drug molecules can be complex, often linking multiple substructures through rotatable bonds. Those bonds let the molecule twist into many three-dimensional shapes, or conformations, making it "floppy" in solution. To bind its target, however, a floppy molecule must lock into a single active shape. Losing that freedom of motion carries an entropic penalty, which weakens binding affinity and lowers potency.

To solve this problem, medicinal chemists often rigidify molecules by adding new bonds. Two substructures linked by a single rotatable bond can twist into many conformations, but adding a second bond to tie them together sharply limits how they can move in three-dimensional space. By placing that bond in the right spot, chemists can lock the molecule into the exact shape needed to bind its target. This strategy, often called "paying the entropic cost in advance," can dramatically boost potency.

The walkthrough focuses on a well-known case of successful rigidification. Crizotinib is a first-generation, small-molecule tyrosine kinase inhibitor (TKI) used to treat certain forms of non-small cell lung carcinoma; it contains five major rotatable bonds. Lorlatinib is a third-generation, small-molecule TKI that resembles crizotinib but has been cyclized at its end to reduce the effective number of rotatable bonds to zero, leading to higher potency among other effects.

Collaborative Intelligence in Action

The DeepMind team asked the agent to search the literature for these drugs and their analogues. With the Cheminformatics Skills in Antigravity, the agent extracted molecules from the documents, associated them with the reported data, and prepared the compounds for downstream analysis.

Within minutes, the agent retrieved thousands of compounds, filtered out irrelevant molecules such as buffers and solvents, and documented their sources in a report. The report also provided an initial analysis of the retrieved compound set, highlighting, for example, that the macrocycles are largely predicted to be brain-penetrant — an important property given that the target lung cancer frequently metastasizes to the brain.

For deeper exploration, the agent launches an interactive viewer displaying the curated compounds, their associated data and customizable scatter plots. Plotting potency against rotatable bonds, colored by macrocycle status, immediately reveals that macrocycles like lorlatinib tend to have fewer rotatable bonds and higher potency.

DeepMind said it designed the skills not only to expand the agent's chemistry capabilities but also to make user-agent collaboration seamless. In the interactive viewer, scientists can browse every compound retrieved from the literature, click individual molecules to inspect their metadata, and plot relationships across the dataset. For example, plotting cellular potency versus rotatable bonds, colored by macrocycle status, clearly shows that the highest-potency compounds are the most rigid.

Matched Molecular Pair Analysis and Predictive Modeling

After downloading and curating all the data, the agent performs more advanced analysis to support the work of a computational chemist. As a next step, the team performed matched molecular pair analysis (MMPA), a technique used in medicinal chemistry to isolate the effects of small, discrete changes to a molecule by comparing pairs of molecules that differ only by that change. Notably, the agent had independently suggested this analysis, as the dataset was expected to include many informative pairs. The team also asked the agent to build predictive models using its molecular machine learning capabilities.

The MMPA compared many pairs of molecules that differ by whether or not they are cyclized. The results show that cyclization reduces the number of rotatable bonds while generally increasing potency, as expected.

The agent also built a large number of potency models using varying feature sets and model architectures, and validated them on a held-out set. The skills provide the agent with guidance on how to best build molecular machine learning models — for example, the agent will often use chronological splits, training on older compounds to predict newer ones, when this information is available, but can fall back to scaffold splits to better estimate generalizability.

In the demonstration, the agent benchmarked multiple machine learning architectures with differing feature sets across four different experimental readouts, using a scaffold split to better gauge out-of-distribution performance. A combination of statistical and visual outputs allows the scientist to assess model performance and then iterate with the agent.

From Analysis to Synthesis Planning

Finally, the team asked the agent to design new, improved macrocycles based on its learnings, and the agent recommended multiple reactions and synthesis routes involving different compounds. The agent can prepare these compounds for further analysis with AlphaFold or physics-based techniques such as molecular dynamics simulation.

The agent is able to reason about chemistry as it proposes new synthesis plans. In one example, it proposed a cyclization reaction, providing the scientist with electron-pushing diagrams to demonstrate the mechanism.

Once the analysis is complete, results can be exported in various formats, ready to be loaded into many industry-standard chemistry software packages.

Knowing the Limits

DeepMind was also explicit about what these tools cannot do. Sparse data, inter-lab assay noise, and narrow dynamic ranges limit the predictive precision of even the most powerful models; computational designs are testable hypotheses, not finished drugs. With the Cheminformatics Skills in Antigravity, a drug discovery scientist can generate more strongly supported design hypotheses faster, but every proposed design must be synthesized and validated experimentally.

The company argued that the speed at which researchers can make meaningful scientific decisions is accelerating. Not long ago, the workflow demonstrated in the post would have taken multiple days and required significant computational expertise. Using Antigravity as a science workbench allows both computational and experimental chemists to accelerate their workflows and focus on the questions that really matter.

Availability and Free Access for 10,000 Scientists

The cheminformatics skills are available on Google Antigravity as of the announcement. Scientists can download the Antigravity App and install the Science plugin to start using the skills, and explore documented Science use cases. Enterprise users can get started with cheminformatics on Antigravity by choosing "Use Business Account" and logging into the app with their Google Cloud credentials.

To further advance the pace of research, Google is offering 10,000 scientists the highest usage tier in Google Antigravity free of charge for 6 months, with onboarding on a rolling basis and subject to eligibility, capacity and selection. Applications are now open.

Acknowledgements

DeepMind acknowledged the contributions of the following team members: Filip Radlinski, Artem Sobolev, Victoria Johnston, Hannah Kiely-Collins, Zachary Wu, Jue Wang, Jack Dunger, Adriana Fernandez Lara, Ngân Vũ, Doug Fritz, Yunhan Xu, Sebastian Nowozin, and Pushmeet Kohli.

The company also pointed readers to related posts on its blog, including "Accelerating the search for antimicrobial peptides with Google Antigravity" and "Conversing with antiquity".

Source: Google DeepMind Blog — "Agentic Cheminformatics with Google Antigravity"