Predicting protein structures used to be a long, resource-heavy process, often requiring local installations or complex data preparation. Today, especially with the introduction of AlphaFold-2, molecular modelers have access to powerful prediction engines remotely. However, integrating these capabilities into real workflows can still be tricky.
If you’re using SAMSON, you can streamline your structure prediction by accessing AlphaFold-2 directly through its Biomolecular Structure Prediction extension. In a few steps, you can go from a FASTA sequence to a colorized 3D model of your protein—no command-line experience required.
Why This Matters
Let’s say you’re working on a newly sequenced protein without any experimental structure. Getting even a basic understanding of its 3D conformation normally means time-consuming modeling pipelines. Now, you can shift from plain sequence to predictive model in just a few clicks. This helps accelerate hypothesis generation, comparative modeling, and even structure-based drug design workflows.
Here’s How to Get Started
- In SAMSON, open Home > Predict.
- Select AlphaFold-2 from the list of available services.
- Upload your FASTA sequence(s).
- Choose the desired AlphaFold model (e.g., monomer or multimer) and a database for the multiple sequence alignment step.
- Click Start prediction.
What Happens Behind the Scenes
Predictions run on cloud servers equipped with high-performance A100 GPUs, and computations are handled remotely. The system securely transmits your data and returns a detailed 3D structure file, including per-residue confidence scores (pLDDT values). In SAMSON, these scores are used to automatically colorize the resultant structure, highlighting reliable regions at a glance.
Costs and Credits
As with all cloud services in SAMSON, AlphaFold-2 predictions use computing credits. Contact the team at contact@samson-connect.net to request credits or buy them online.
Good Practices
- Ensure your FASTA sequence is clean and properly formatted before uploading.
- If working on protein complexes, select the multimer model to capture inter-chain interactions.
- Download and archive your prediction results for future reference or publication.
- Whenever AlphaFold predictions contribute to your research output, include the appropriate citations, such as Jumper et al. (2021).
Preview Before Sharing
All prediction results can be reviewed through the Interface > Cloud jobs section or your SAMSON Connect user account. Ready-to-share visuals make collaboration with colleagues easier and more insightful.
Learn more in the full documentation.
SAMSON and all SAMSON Extensions are free for non-commercial use. You can download SAMSON at https://www.samson-connect.net.