A prepared receptor pocket, target context, candidate count, fragment or scaffold constraints, and design objectives.
Capability 09
Molecule generation
Which new molecular candidates satisfy the stated pocket, fragment, and design constraints?
What FROGENT does
Executes a four-stage design–build process: generate candidates, apply constraints, refine and diversify, then export the candidate structures.
Candidate molecular structures represented as SMILES with generation artifacts and constraint context.
Orchestrate · Forge
Design workflowManuscript workflow figure
Read the task from question to recorded output.
Read the figure left to right through four design stages. Each stage narrows or reshapes the candidate set while preserving the original task constraints in campaign state.
Molecule Generation Workflow. Forge generates initial candidates, constrains them with known fragments, refines and diversifies the set, and exports final candidate SMILES.
Download source figure (PDF)What the evidence supports
Interpret the workflow at the level the manuscript shows.
The manuscript evaluates de novo design using CrossDocked protein pockets and uses molecular generation in the cardiomegaly and carbonic anhydrase case studies.