This repository was archived by the owner on Mar 10, 2026. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 0
This repository was archived by the owner on Mar 10, 2026. It is now read-only.
Implement Standard Placement Workflow #18
Copy link
Copy link
Description
Description:
This task focuses on implementing the core scientific capability of SCREAM++: the standard side-chain placement workflow. This workflow will take a molecular system and a configuration, and produce one or more optimized structures with predicted side-chain conformations.
Tasks:
- Phase 0: Setup: Load resources (
Forcefield,RotamerLibrary), create theOptimizationContext, and resolve the set of active residues to be optimized. - Phase 1: Pre-computation: Invoke
tasks::el_energy::runto compute and cache the Empty Lattice (EL) energies for all rotamers of the active residues. - Phase 2: Initialization: Create an initial
working_systemby placing the ground-state rotamer (lowest EL energy) for each active residue. Initialize theOptimizationStatewith this first conformation. - Phase 3: Clash-Driven Optimization: Implement an iterative loop that repeatedly:
- Identifies the most severe steric clash in the system using
tasks::clash_detection. - Resolves the clash by performing a local optimization on the clashing pair using
tasks::doublet_optimization. - Updates the
working_systemandOptimizationState. - Continues until no significant clashes remain or the system's energy converges.
- Identifies the most severe steric clash in the system using
- Phase 4: Global Search (Simulated Annealing - Optional): Implement an optional Monte Carlo simulated annealing loop based on the
simulated_annealingconfiguration to escape local minima. - Phase 5: Final Refinement: Implement a final, greedy iterative refinement loop (singlet optimization) where each active residue is optimized one-by-one in the context of all others, ensuring the system settles into a local energy minimum.
- Phase 6: Finalization: Extract the top N best-scoring solutions from the
OptimizationStateand return them.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
enhancement ✨New feature or requestNew feature or request