What is Neo-1
Neo-1 is the world's first AI model launched by VantAI that unifies de novo molecule generation with atomic-level structure prediction. It can predict biomolecular structures, generate entirely new molecules, and excels in designing novel therapeutic drugs such as molecular glues. Neo-1's multimodal input capability can accept various information such as partial sequences, partial structures, and experimental data, significantly improving the efficiency and accuracy of drug design. Combined with VantAI's NeoLink platform, Neo-1 generates sparse structural constraints through crosslinking mass spectrometry, then assembles them into complete atomic-resolution structures, advancing the development of structural biology.
Neo-1's main features
- Unified generation and prediction: Neo-1 is the first model to unify de novo molecule generation with atomic-level structure prediction. By generating latent representations of molecules rather than predicting atomic coordinates, it can predict biomolecular structures and generate entirely new molecules.
- Multimodal input: Neo-1 accepts input from multiple modalities, including partial sequences, partial structures, and experimental data. The multimodal input approach significantly improves the model's flexibility and applicability.
- Large-scale training: Neo-1 is one of the largest diffusion-based models in biology, trained on structural and synthetic datasets using hundreds of NVIDIA H100 GPUs.
- Custom datasets and tools: Neo-1 combines VantAI's proprietary NeoLink dataset with the PINDER & PLINDER tools co-developed with NVIDIA.
Neo-1's technical principles
- Diffusion process in latent space: Neo-1 shifts the diffusion process from traditional coordinate space to latent space. This shift enables the model to reason in smoother sequence and structure landscapes, generating entirely new molecules, including proteins, peptides, and small molecules, while predicting their structures with atomic-level precision.
- Large-scale training and custom datasets: Neo-1 is one of the largest diffusion-based models in biology, trained on structural and synthetic datasets using hundreds of NVIDIA H100 GPUs. It combines VantAI's proprietary NeoLink dataset with the PINDER & PLINDER tools co-developed with NVIDIA, improving the model's performance.
- Precise molecule generation and structure prediction: Neo-1 adopts a 'coarse-to-fine' generation approach, capable of applying intermediate rewards based on the entire molecular structure to guide molecule generation toward any target. Unlike traditional autoregressive models, which lack flexibility during the generation process.
Neo-1's project address
- Project official website：https://www.vant.ai/neo-1
Neo-1's application scenarios
- Molecular glue design: Neo-1 can design novel therapeutic drugs such as molecular glues for complex targets, shortening what traditionally takes years to weeks.
- Protein complex structure prediction: Neo-1 can predict the structures of various biomolecular complexes, including ternary complexes, antibody-antigen interactions, and protein-peptide complexes.
- Application of the NeoLink data platform: Combined with VantAI's NeoLink data platform, Neo-1 can assemble complete atomic-resolution structures based on sparse structural constraints generated by crosslinking mass spectrometry.
- Antibody discovery: Neo-1 enables end-to-end rational antibody discovery. It can take partial antibody sequences and antigen structures as input, while folding VH antibody fragments and generating partial CDRH3 sequences.