Shape-Comp

Workflow recap

• Choose a PDB or mmCIF complex from an upload, dataset, or previous job output.
• Set Chain Group A and optionally Chain Group B. If Group B is empty, all remaining chains are used.
• Outputs are written to /outputs as shape_comp.json and shape_comp.tsv, or with your chosen result name.
• Batch mode scores every supported structure in a selected folder and writes a summary JSON/TSV pair.

Example arguments

--input=/inputs/complex.pdb
--chains-a=B
--chains-b=A

--batch-folder=/inputs/designs
--chains-a=B
--chains-b=A
--recursive
--continue-on-error

Options

• --chains-a=A,B: required first molecule or chain group.
• --chains-b=C: optional second group; omitted means all remaining chains.
• --model=1: 1-based model index for multi-model structures.
• --altloc-policy=fail|highest_occupancy|sc_rs: alternate-location handling.
• --include-hetatm and --include-hydrogens: include records excluded by default.
• --output-prefix=shape_comp: basename for JSON and TSV outputs.

Notes

• Supported extensions: .pdb, .ent, .cif, .mmcif.
• No GPU is allocated; CPU limit is 4 vCPUs with 8 GiB RAM.
• This is a geometric interface metric, not a Rosetta score or an energy model.