DCFold: Efficient Protein Structure Generation with Single Forward Pass
For researchers and practitioners in protein design and virtual screening, DCFold dramatically reduces inference time while maintaining state-of-the-art accuracy, enabling practical deployment of high-quality protein structure generation.
DCFold achieves AlphaFold3-level accuracy in protein structure generation with a 15x inference speedup by using a single-step generative model with a Dual Consistency training framework and Temporal Geodesic Matching scheduler.
AlphaFold3 introduces a diffusion-based architecture that elevates protein structure prediction to all-atom resolution with improved accuracy. This state-of-the-art performance has established AlphaFold3 as a foundation model for diverse generation and design tasks. However, its iterative design substantially increases inference time, limiting practical deployment in downstream settings such as virtual screening and protein design. We propose DCFold, a single-step generative model that attains AlphaFold3-level accuracy. Our Dual Consistency training framework, which incorporates a novel Temporal Geodesic Matching (TGM) scheduler, enables DCFold to achieve a 15x acceleration in inference while maintaining predictive fidelity. We validate its effectiveness across both structure prediction and binder design benchmarks.