OpenFold3 on Amazon SageMaker JumpStart: Predict biomolecular structures
Deploy OpenFold3-preview from Amazon SageMaker JumpStart to predict 3D structures of proteins, DNA, RNA, and ligand complexes in your own AWS account. The post covers asynchronous inference, multi-GPU parallelism, bringing your own MSAs for accuracy and privacy, and scaling to zero to reduce cost.
Amazon SageMaker JumpStart now offers the OpenFold3 model, an all-atom structure prediction tool for biomolecules. This open-source model, developed by the OpenFold Consortium and the AlQuraishi Lab at Columbia University, can predict the 3D structures of proteins, DNA, RNA, small-molecule ligands, and their complexes. Users can easily deploy OpenFold3 on Amazon SageMaker using a few lines of code and predict structures with confidence scores.
The endpoint provides network isolation for secure storage of sequences and supports parallel predictions on multi-GPU instances. OpenFold3 is licensed under Apache 2.0, allowing researchers and pharmaceutical companies to inspect, fine-tune, and build upon the model. It runs in a SageMaker managed container with GPU warmup and request queuing, automatically handling autoscaling.
The model predicts structures in mmCIF format, with metrics like pLDDT, PAE, and PDE. Asynchronous inference allows submitting requests to Amazon S3 and processing them over a 1-hour timeout period, useful for high-throughput screening pipelines. Prerequisites include an AWS account, SageMaker AI execution role, and sufficient quota for ML endpoint instances.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
