Predicting the 3D structure of proteins is a fundamental task in structural biology and molecular modeling. However, even with state-of-the-art deep learning systems like AlphaFold-2, understanding how to access and use these tools efficiently can be a barrier for many researchers. If you’re using or exploring SAMSON, a molecular design platform, this post walks you through how AlphaFold-2 structure predictions are integrated directly into your workflow—without needing to install complex dependencies or set up local computing environments.
Why Is This Helpful?
Setting up AlphaFold-2 locally is resource-intensive: it requires significant storage for sequence databases, powerful GPUs for efficient runs, and technical know-how for managing execution environments. SAMSON integrates AlphaFold-2 as a Cloud-based service via its Biomolecular Structure Prediction Extension, so you can bypass the complexity and get accurate predictions directly from your molecular modeling workspace.
How It Works in SAMSON
Here’s how to begin predicting structures using AlphaFold-2 directly in SAMSON:
- Navigate to Home > Predict in the SAMSON interface.
- Select the AlphaFold-2 service.
- Upload one or more FASTA files containing your protein sequences.
- Choose the appropriate model variant (e.g., monomer, multimer) and the database for multiple sequence alignment.
- Click Start prediction to send your job to the Cloud.
That’s it. No command-line instructions, manual file management, or infrastructure setup required.
What Happens Behind the Scenes
Your prediction is sent to secured Cloud resources, including high-performance instances with A100 GPUs. Depending on your model and inputs, prediction time and credit usage may vary. Prediction results are stored as Cloud jobs, viewable in:
- Interface > Cloud jobs (within SAMSON)
- SAMSON Connect > Account > Jobs
Once completed, the predicted structure is automatically colorized by pLDDT confidence values (when available), helping you visually interpret areas of high or low structural reliability.
Cost and Accessibility
Predictions consume computing credits, which can be purchased directly through SAMSON Connect or requested via email at contact@samson-connect.net. Pricing varies depending on the compute resources selected. This enables flexible usage based on your modeling needs, from individual structures to larger batch runs.
Publication Reminder
If you use AlphaFold-2 prediction results in scientific publications, you’re encouraged to cite the original paper: Jumper, J. et al., Nature 596, 583–589 (2021). For multimer mode, please also check if an additional citation is needed.
A Quick Tip
The simplicity of the workflow makes sharing results easy. You can distribute prediction results to colleagues via exported data files or cloud links available through your SAMSON Connect account, all without ever leaving the application.
Learn More
To dive deeper into the different structure prediction services available in SAMSON, including Boltz-2 and Chai-1, please visit the full documentation page: https://documentation.samson-connect.net/tutorials/bsp/bsp/.
SAMSON and all SAMSON Extensions are free for non-commercial use. You can get SAMSON at https://www.samson-connect.net.