AlphaFast: High-throughput AlphaFold 3 via GPU-accelerated MSA construction
Benjamin Perry ⋅ Jeonghyeon Kim ⋅ Philip Romero
Abstract
AlphaFold 3 (AF3) enables accurate biomolecular modeling but is limited by slow, CPU-bound multiple sequence alignment (MSA) generation. We introduce AlphaFast, a drop-in framework that integrates GPU-accelerated MMseqs2 sequence search to remove this bottleneck. AlphaFast achieves a 68.5× speedup in MSA construction and a 22.8× reduction in end-to-end runtime on a single GPU, and delivers predictions in 8 seconds per input on four GPUs while maintaining indistinguishable structural accuracy. A serverless deployment enables structure prediction for as little as \$0.035 per input. Code is available at https://github.com/RomeroLab/alphafast
Chat is not available.
Successful Page Load