protenix

You can predict protein structures and complexes by providing a valid input JSON detailing sequences and seeds. The automation processes single chains, multi-chain complexes, and ligand-inclusive systems, emitting confidence metrics like pLDDT, ipTM, pTM, and ranking_score for validation.

Assessing complex pose stability requires running a multi-seed ensemble capability. By evaluating predictions across multiple seeds and samples, you can measure structural stability and facilitate robust decision-making for downstream analysis.

Yes, structure prediction applies to single chains, multi-chain complexes, and protein-ligand systems. It processes sequence inputs to generate predicted structures and emits interpretable confidence metrics like ipTM and ranking_score for complex pose assessment.

Yes, protein structure prediction requires a Protenix installation on PATH. You must also set the environment variables PROTENIX_ROOT_DIR or PROTEUS_FOLD_DIR to correctly locate model assets before running the prediction wrapper.

Input validation for protein complex prediction enforces a specific JSON specification. An input validation wrapper checks your sequences and seeds against the spec, supporting on-demand compute routing via a Python entry point before executing the prediction.