FoldSense uses pyPept for the BILN-to-chemical-graph/generated-3D representation lane. Each judge protein must have its own generated PDB and standalone 3D view before the pyPept gate is PASS. These outputs are generated conformers—not experimental structures and not molecular-dynamics trajectories.
The generation manifest retains pyPept version, upstream commit, input BILN, return code, exact PDB/PNG/HTML hashes, byte counts and failed entries. Failure is not replaced with an experimental PDB.
Loading /structures/manifest.json…
The main Molecular Reader can switch between a pyPept-generated fragment and a cited experimental PDB context. The UI labels the active source explicitly; no generated conformer is promoted to experimental evidence.