# Structure Provenance

This project now has a machine-readable structure layer in `src/model/structures.js` and a JSON mirror in `data/structures.json`.

The current rule is strict:

```text
PDB-backed visual anchor does not mean atomistic simulation.
```

Where coordinates have not yet been parsed, the UI says so.

## Current Structure Anchors

| PDB | Component | Organism | Resolution | Current Use | Status |
| --- | --- | --- | ---: | --- | --- |
| 5XTD | Complex I | Homo sapiens | 3.7 A | Complex I L-shape, FMN/Fe-S/Q candidate geometry source | visual anchor; cofactor distances pending |
| 8Q1U | Complex I Q-site comparative anchor | Bos taurus | 3.3 A | Comparative mammalian N2-to-ubiquinone geometry | Q-bound comparative anchor |
| 5XTE | Complex III | Homo sapiens | 3.4 A | Complex III/Q-cycle and Qo-site ROS context | visual/provenance anchor |
| 5XTH | Supercomplex I1III2IV1 | Homo sapiens / Bos taurus context | 3.9 A | Respirasome-scale organization | membrane/supercomplex anchor |
| 5XTI | Megacomplex I2III2IV2 | Homo sapiens / Bos taurus context | 17.4 A | High-order respiratory assembly | architecture anchor only |
| 6TT7 | ATP synthase | Ovis aries | 3.5 A | Mammalian Complex V/proton-return anchor | Complex V proxy |

## What Is Structural Now

- Structure IDs, titles, organisms, methods, and resolutions are explicit.
- Each structural source has a declared use and caveat.
- The protein zoom level displays the active structural anchor.
- The Complex I cofactor path is labeled as a reduced educational network.
- `5XTD` supplies human Complex I FMN/Fe-S geometry.
- `8Q1U` supplies a comparative bovine Q-bound Complex I edge.
- Extracted distances now use redox-atom filters where configured.

## What Is Not Structural Yet

- Fe-S center names are not fully manually verified.
- The `8Q1U` Q edge is headgroup-filtered but may reflect open/deactivated-state geometry.
- Protein surface shapes are not atomistic meshes.
- The app does not yet load EM maps, density, or full atomic models.

## Next Upgrade

The extraction script now exists. The next serious step is curation:

```text
verify Fe-S center residue IDs -> validate Q pose/conformation
-> compute per-hop rates -> render hop-specific path kinetics
```

For Complex I, the first target is:

```text
FMN -> N3 -> N1b -> N4 -> N5 -> N6a -> N6b -> N2 -> Q/Q-site
```

Once that exists, the tunneling window can display hop-specific distances and rates rather than one effective distance.