If you’re a molecular modeler working with proteins and struggling to generate accurate 3D structures from FASTA sequences, you’re not alone. Waiting for software installations, configuring environments, and handling compatibility issues can be a huge time sink. Fortunately, SAMSON offers a streamlined alternative: cloud-based structure prediction using AlphaFold-2, directly within its interface.
Let’s walk through why and how to use AlphaFold-2 in SAMSON—and why it can save you time and effort when predicting biomolecular structures.
Why AlphaFold-2 in SAMSON?
AlphaFold-2 revolutionized protein structure prediction by achieving near-experimental accuracy for many proteins based solely on their amino acid sequences. But setting it up locally is complex. Running it in SAMSON, on the other hand, requires no installation—just access the service in SAMSON’s Predict menu and send your job to a secured cloud environment with high-performance GPUs (like A100 instances).
This not only simplifies your workflow but also ensures you’re using the most up-to-date models without stressing your local machine.
Getting Started with AlphaFold-2
Here’s how to run AlphaFold-2 predictions in SAMSON:
- Open SAMSON and go to Home > Predict.
- Select the AlphaFold-2 service.
- Upload one or more FASTA files.
- Choose a prediction model: monomer, multimer, etc.
- Select a database for the multiple sequence alignment (MSA)—a critical step for quality predictions.
- Click Start prediction.
You can monitor the progress from Interface > Cloud jobs in SAMSON, or check it online in SAMSON Connect > Account > Jobs.
Color and Confidence: Interpreting the Results
Once your predicted structure is loaded, SAMSON colorizes the protein based on pLDDT scores (predicted Local Distance Difference Test)—a measure of confidence for each residue. Higher pLDDT values (typically shown as blue) indicate areas with higher structural confidence, while lower values (yellow or orange) may point to flexible or uncertain regions.
This visual feedback is instant and can guide decisions like model refinement or experimental validation.
Computing Resources and Cost
Predictions are run using powerful cloud machines, including A100 GPUs, which significantly reduce computation time. Jobs require computing credits, and users can either request them or buy credits directly. This pay-as-you-go model ensures you only pay for what you need.
Publishing? Don’t Forget to Cite
If you use results from AlphaFold-2 predictions in publications, make sure to cite the original AlphaFold-2 paper. If you use multimer models, refer to the AlphaFold-Multimer publication as well.
SAMSON’s integration with AlphaFold-2 offers a no-hassle, visual-first approach to structural biology. Whether you’re preparing figures, analyzing mutations, or building models for simulation, it helps streamline a critical step in your research pipeline.
📄 Read the full documentation here to learn more.
SAMSON and all SAMSON Extensions are free for non-commercial use. You can download SAMSON here.