Oriel — AI & technology worth noticing, every day.

OpenFold3 on Amazon SageMaker JumpStart: Predict biomolecular structures

Web page · https://aws.amazon.com/blogs/machine-learning/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.

First seen by Oriel on .

AI summary

This AWS post demonstrates deploying OpenFold3-preview through SageMaker JumpStart to asynchronously predict all-atom structures of proteins, nucleic acids, ligands, and complexes within an AWS account.

  • OpenFold3 0.4.1 uses the of3-p2-155k checkpoint, accepts sequences or ligand SMILES, and returns mmCIF structures with pLDDT, PAE, and PDE confidence metrics.
  • SageMaker Asynchronous Inference stages requests in Amazon S3, queues jobs for up to one hour, and writes results or errors to configured S3 locations.
  • Each query produces five diffusion-sampled candidate structures by default, with aggregated confidence files containing sample_ranking_score, avg_plddt, pTM, and ipTM.
  • Supported instances include ml.g6.2xlarge, ml.g6.4xlarge, and ml.g6.12xlarge, while asynchronous scaling can queue batches and scale to zero when idle.
  • The endpoint uses network isolation, and users can protect sequence and structure data with dedicated S3 permissions, encryption, and lifecycle policies.

#openfold3 #sagemaker #biomolecular structure prediction #asynchronous inference

Generated by gpt-5.6-sol. Check the original for details.

Aliases

SourceExternal IDURL
AWS Machine Learning Blog4dea8ce77a40d45986adf65199b169c60e84bcc1https://aws.amazon.com/blogs/machine-learning/openfold3-on-amazon-sagemaker-jumpstart-predict-biomolecular-structures/ Read discussion