Applications of modular co-design for <i>de novo</i> 3D molecule generation

2026/01/01 by Danny Reidenbach, Filipp Nikitin, Olexandr Isayev +1 · 1 voice
Chemistry · Materials Science · Biochemistry, Genetics and Molecular Biology · #Crystallography and molecular interactions #Supramolecular Self-Assembly in Materials #Chemical Synthesis and Analysis

paper · pdf · doi:10.1039/d5dd00380f

openalex publication_date 2026/01/01 · openalex created_date 2026/01/09 · openalex updated_date 2026/06/22

Abstract

A scalable and accurate transformer model unifies diffusion and flow-matching approaches to generate realistic 3D molecules.

Citations

• Merck molecular force field. I. Basis, form, scope, parameterization, and performance of MMFF94
• Los carnavales de La Bañeza: Aspectos antropológicos
• Denoising Diffusion Probabilistic Models
• PROTAC targeted protein degraders: the past is prologue
• Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioning
• Structure-based Drug Design with Equivariant Diffusion Models
• Scalable Diffusion Models with Transformers
• Improving and generalizing flow-based generative models with minibatch optimal transport
• Evolutionary-scale prediction of atomic-level protein structure with a language model
• Stochastic Interpolants: A Unifying Framework for Flows and Diffusions
• 3D molecule generation by denoising voxel grids
• Harmonic Self-Conditioned Flow Matching for Multi-Ligand Docking and Binding Site Design
• SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformers
• Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design
• CREST—A program for the exploration of low-energy molecular chemical space
• Accurate structure prediction of biomolecular interactions with AlphaFold 3
• SemlaFlow -- Efficient 3D Molecular Generation with Latent Attention and Equivariant Flow Matching
• Simulating 500 million years of evolution with a language model

Discussions

• 👉 Don’t miss this Advance Article from #DigitalDiscovery featuring Megalodon 🦈, a scalable transformer-based architecture for multi-modal molecular diffusion and flow matching towards de novo 3D-mol [bsky, 0 points, 0 comments]