Domestic Generative AI Model for Antibody Design Matches AlphaFold3, Predicts Antigen-Antibody Complex Structures and Enables De Novo Antibody Design

Amid the wave sparked by AlphaFold 3, a new generative AI large model for antibody design has emerged.

Named GeoFlow, it can be used simultaneously for antigen-antibody complex structure prediction and de novo antibody design.

For example, given an antigen structure and a specific epitope, GeoFlow can generate entirely new antibody molecules:

△ Schematic diagram of de novo antibody generation based on GeoFlow

In the task of antigen-antibody complex structure prediction, on a test set consisting of 66 antigen-antibody complex structures, GeoFlow achieved a Top-1 success rate of 43.9%, matching AlphaFold 3.

The GeoFlow R&D team comes from Biogeo, a generative AI-driven protein design and development platform company. Biogeo was founded in 2022 by Dr. Jian Tang, an AI drug discovery scientist, with Yoshua Bengio, one of the “Three Giants” of AI and Turing Award laureate, serving as Chief Scientific Advisor.

What Does GeoFlow Look Like?

Recently, AlphaFold 3, developed by teams including Google DeepMind and Isomorphic Labs, was published in Nature, drawing widespread attention from the industry.

Compared to previous-generation methods, AlphaFold 3 has expanded its prediction scope to almost all biological molecules and their interactions, marking another important milestone for generative AI in the life sciences.

One of the model’s key innovations is the use of a popular generative AI technique—the diffusion model (whereas AlphaFold 2 was a discriminative AI model)—to directly generate the 3D coordinates of each atom.

If traditional discriminative AI is like a music critic who can identify and evaluate a song’s rhythm, style, and arrangement, then generative AI is like a singer who can create new works.

From evaluating data to generating data, the application scenarios for AI have been greatly expanded. For instance, in structure prediction tasks, generative AI can sample more conformations faster; in protein design tasks, it can explore protein space more efficiently to design complex protein molecules with intended functions.

Diffusion generative models were initially used primarily for image generation (recently also applied to 3D video generation, such as Sora).

The core team at Biogeo began applying diffusion models to the 3D structure generation of molecules in 2021. Their paper on GeoDiff was among the top 50 most cited papers in AI in 2022.

Building on this technical foundation, they have developed the latest generative AI large model for antibody design: GeoFlow.

The GeoFlow model architecture is as follows:

GeoFlow is based on a geometric deep learning architecture and the latest flow matching generative model. It can be used simultaneously for:

• Antigen-antibody complex structure prediction: Inputting antigen structures/sequences and complete antibody sequences, the model generates the antigen-antibody complex structure.
• Antibody design: Inputting antigen structures and antibody sequences, with the CDR regions to be designed represented as masks, the model generates the complex structure and the CDR region sequences.

Modeling antigen-antibody interaction forces at the atomic level is the core difficulty in these two tasks.

Unlike existing Transformer architectures, GeoFlow adopts a geometric deep learning foundation model, which better models atom-to-atom relationships in three-dimensional space.

Regarding generative model selection, GeoFlow employs the latest flow matching model. Compared to diffusion generative models, flow matching models are more efficient and robust during both training and inference.

Antigen-Antibody Complex Structure Prediction Rivals AF3

The research team evaluated GeoFlow’s performance on the antigen-antibody complex structure prediction task.

Antigen-antibody complex structure prediction plays a crucial role in antibody drug discovery. However, accuracy remains suboptimal for both traditional methods based on energy functions (such as HDock and MOE) and deep learning-based prediction models (such as AlphaFold 2 Multimer).

On a test set consisting of 66 antigen-antibody complex structures (published after 2023), GeoFlow’s Top-1 success rate (defined as successful if the DockQ score of the structure with the highest model ranking is “Acceptable” or above) reached 43.9%, matching AlphaFold 3 and approximately doubling that of AlphaFold 2 Multimer.

△ Evaluation results for antigen-antibody complex prediction

Although traditional molecular docking methods can also generate multiple possible structures, their scoring accuracy is low, limiting their practical application value.

△ Comparison of prediction results by various models for PDB 8BLQ (left) and 8DOK (right)

Compared to AlphaFold 3, GeoFlow can not only be used for antigen-antibody complex structure prediction but also for de novo antibody design and optimization.

For traditional AI methods, de novo design of large molecules is very difficult. The main reason is the difficulty in quickly sampling high-quality samples; discriminative models must evaluate a vast number of low-quality samples from an immense molecular space, akin to finding a needle in a haystack.

The emergence of generative AI has brought revolutionary opportunities for large molecule design.

Taking the HER2 target as an example, based on the binding epitope of the approved antibody drug Herceptin, the research team used GeoFlow to generate a small antibody library, which was then screened using a phage display library. Among the ten candidate sequences obtained:

• Binding Activity: Six molecules showed binding comparable to Herceptin in ELISA experiments, reaching nanomolar levels. BLI results indicated that the affinity of molecule #1 and #3 was even 2-3 times higher than that of Herceptin.
• Binding Epitope: Competitive ELISA showed strong competition between these six molecules’ binding and Herceptin, suggesting their binding epitopes are consistent with those of Herceptin.

These results demonstrate the application prospects of generative AI in the field of de novo large molecule design.

About Biogeo

Biogeo is a generative AI-driven protein design and development platform company founded by AI scientist Dr. Jian Tang in 2022, with Yoshua Bengio, known as the “Father of AI” and Turing Award laureate, serving as Chief Scientific Advisor.

The company’s business focuses on building AIGC large models to understand the language of life, creating multimodal large models that bridge natural language and protein language, reconstructing the process of antibody drug discovery and design, and developing programmable proteins for applications in biomedicine and bio-manufacturing.

Biogeo’s generative AI large models currently cover stages such as large molecule design, screening, and modification, and have developed GeoBiologics, a one-stop antibody discovery platform.

The GeoFlow model is currently open for non-commercial testing of antigen-antibody complex structure prediction, supporting eight predictions per week with an input limit of 1,150 amino acids per task.

Test address: https://geobiologics-lite.biogeom.com/about