A structure prediction can stall before it begins—not because the biology is unusually complicated, but because the input is not ready. A sequence may be in a FASTA file, a ligand may be described by a SMILES string, or a prediction may need a specific modification or distance restraint. Choosing a service that accepts the system you want to model is a practical first step toward avoiding unnecessary setup and reruns.
SAMSON’s Biomolecular Structure Prediction extension provides access to AlphaFold-2, Boltz-2, and Chai-1 through Home > Predict. The useful way to choose between them is to start with your input and the options you need—not with the assumption that one service is right for every job.
Start with the material you have
- Protein FASTA files: AlphaFold-2 accepts one or more FASTA files. During setup, you can choose an AlphaFold model, such as monomer or multimer, and a database for multiple sequence alignment.
- Protein, DNA, RNA, or ligand inputs: Boltz-2 lets you add these components individually. Enter sequences for the biological polymers, CCD codes where applicable, and SMILES strings for ligands.
- Mixed components or distance restraints: Chai-1 also lets you add proteins, DNA, RNA, and ligands. Its setup includes optional pocket restraints and contact restraints.
This quick inventory helps turn “Which prediction should I run?” into a more answerable question: “Which service accepts the components and setup I need?” If you are unsure about the input format, prepare the sequence, file, code, or SMILES string before opening the prediction interface.
When extra setup matters
For AlphaFold-2, the model choice and multiple-sequence-alignment database are explicit parts of the setup. Confirm that the selected model matches the system you intend to submit, then add the required FASTA file or files.
Boltz-2 supports adding a modification to a protein, DNA, or RNA sequence. The interface asks for the residue index and the modification’s CCD code. That makes it especially important to check the position against the sequence you are submitting: a correct modification at the wrong index is still the wrong input.
With Chai-1, restraints are optional. A pocket restraint sets a distance restraint between a residue and a chain; a contact restraint sets one between two residues. Add them when they express a relationship you want included in the prediction setup. Before submitting, check that you have selected the intended residue or chain for each restraint.
Plan for cloud jobs and credits
Predictions run in the cloud, and data are sent via a secure connection. You can follow results in SAMSON under Interface > Cloud jobs or on SAMSON Connect > Account > Jobs.
Costs depend on the service. The documentation notes that AlphaFold-2 predictions can use cloud machines with varying performance and costs, including instances with A100 GPUs. Boltz-2 and Chai-1 use a dedicated A100 GPU and computing credits; a typical prediction with either is listed at 0.5–1 credit. Check your available credits before submitting, especially if you are testing multiple input variants. You can buy computing credits or contact SAMSON support to request them.
A short preflight checklist
- Have the correct sequence files, sequences, CCD codes, or SMILES strings ready.
- Confirm the service accepts the components you want to submit.
- Check model and alignment-database choices for AlphaFold-2.
- For Boltz-2 modifications, verify the residue index and CCD code.
- For Chai-1 restraints, verify the residues, chain, and intended distance relationship.
- Check computing-credit availability, then submit and monitor the job.
When the structure is loaded, SAMSON can colorize it by pLDDT values if those values are available in the result files. Treat this as a way to inspect the prediction, not as a substitute for deciding whether the model fits your research question. Depending on your next step, you can compare candidates, prepare a selected structure for docking or simulation, or continue with visualization.
For the full service instructions and related tutorials, read the biomolecular structure prediction documentation.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON at samson-connect.net.