A structure-prediction job can stall before it even starts: the system contains more than a protein sequence. Perhaps there is a ligand to describe, a modified residue to include, or nucleic acid alongside the protein. Choosing a service that accepts the inputs you have—and preparing those inputs before opening the prediction tool—can make the workflow much clearer.
SAMSON’s Biomolecular Structure Prediction extension provides access to AlphaFold-2, Boltz-2, and Chai-1. The useful first question is not simply “Which model is best?” It is “What do I need to provide for this particular system?” The available inputs and options differ across the three services, so matching the workflow to the system is a practical way to avoid unnecessary setup.
Match the input to the service
AlphaFold-2 is launched with one or more FASTA files. In the prediction panel, you choose a model, such as monomer or multimer, and a database for multiple sequence alignment. This is a straightforward route when your starting material is sequence information in FASTA format.
Boltz-2 lets you build up a system by adding proteins, DNA, RNA, or ligands. Enter the corresponding sequences, CCD codes, or ligand SMILES strings. If you need to represent a modification in a protein, DNA, or RNA sequence, the interface also lets you add a modification by specifying its residue index and CCD code.
Chai-1 also accepts proteins, DNA, RNA, and ligands, entered as sequences or SMILES strings. Its workflow includes optional restraints: a pocket restraint can specify a distance relationship between a residue and a chain, while a contact restraint can specify a distance relationship between two residues.
These differences suggest a simple planning habit: write down the components of your system and the form in which you have them before selecting a service. For example, note whether you have FASTA files, ligand SMILES, CCD codes, or a modification to represent. If a restraint matters to your setup, check whether the service’s available restraint options fit what you intend to specify. This small inventory can prevent a mismatch between your input and the interface.
A short preflight checklist
- List every component. Include protein, DNA, RNA, and ligand components as relevant.
- Check the input format. Have sequences, FASTA files, CCD codes, or SMILES strings ready, depending on the service.
- Record special details. For Boltz-2 modifications, identify the residue index and CCD code. For Chai-1, decide whether you need a pocket or contact restraint.
- Consider the job cost. Boltz-2 and Chai-1 predictions use cloud computing credits; a typical prediction for either is listed as costing between 0.5 and 1 credit.
- Know where to find the result. Results can be viewed in SAMSON under Interface > Cloud jobs, or on SAMSON Connect under Account > Jobs.
After the prediction finishes
Loading a result into SAMSON gives you a place to inspect the structure and compare candidates when needed. If pLDDT values are present in the result files, SAMSON colorizes the structures based on those values. Treat this as a visual aid for examining the prediction, not as a replacement for checking whether the model is suitable for your next step.
Once you have selected a model, the next workflow depends on your goal. You might inspect it, compare it with a known structure, or prepare it for docking or simulation. The documentation also points to tutorials on protein preparation, alignment, and GROMACS workflows for those follow-on tasks.
Predictions are performed in the cloud over a secure connection. For AlphaFold-2, the available cloud machines have varying performance and costs; the documentation notes that options include instances with A100 GPUs. Boltz-2 and Chai-1 are described as running on a dedicated A100 GPU and using computing credits. Check your available credits before launching jobs, especially when planning multiple predictions.
For the complete steps and service details, read the biomolecular structure prediction tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.