H-NS Filament Minimal Dataset

This folder contains a compact, package-friendly subset of the full filament MD dataset.

Purpose

The goal is to avoid shipping the full ~1.7 GB trajectory bundle while preserving deterministic filament construction behavior for the tutorial's non-random assembly path.

This minimal dataset is designed for:

• Assembler.load_minimal_site_map(...)
• assembler.add_dimer(segment='minimal')

Contents

• s1s1_start.pdb
• s2s2_start.pdb
• s1s1_extend.pdb
• s2s2_extend.pdb
• complex_frame_1.pdb
• manifest.json

Size

• Total: ~1.5M

How it works

Unlike the full dataset (many .xtc trajectories), this folder stores only a few selected source-state PDB frames:

• start-state source frames for s1s1 and s2s2
• extend-state source frames for s1s1 and s2s2
• one DBD-DNA complex frame

At runtime, Assembler.load_minimal_site_map(...):

1. Loads these source PDBs
2. Rebuilds segment/site maps using SiteMapper
3. Selects the required site structures for start/extend assembly
4. Uses complex_frame_1.pdb for add_dna(...)

So the segmentation logic is still performed in code (like full mode), but from minimal source inputs.

Generation

This dataset is generated from examples/data/filament_dataset with:

python examples/scripts/extract_minimal_filament_dataset.py --output-dir examples/data/filament_minimal

The selection metadata (trajectory ids, stride, source mapping) is stored in manifest.json.

Notes

• segment='minimal' is intended for reproducible compact usage.
• segment='fixed' and segment='random' remain tied to full trajectory-style site maps.
• If you regenerate with a different --dna-frame-idx, the output complex file name and manifest should stay consistent for your workflow.