A-CODE model achieves fully atomic protein co-design with unified diffusion

Researchers have introduced A-CODE, a novel unified diffusion model for protein co-design that simultaneously optimizes atomic types and coordinates. This one-stage approach differs from traditional two-stage methods by inferring residue identities directly from atom-level predictions. A-CODE demonstrates superior performance in unconditional protein generation and significantly improves success rates in binder design tasks, outperforming existing models. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces a new generative modeling framework for biomolecular systems, potentially accelerating drug discovery and protein engineering.

RANK_REASON This is a research paper detailing a new model for protein co-design. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• A-CODE
• arXiv

• paper
• model release