## Drosophila NeuroXiv data and source

Drosophila NeuroXiv explores 139,273 FlyWire neurons in 78 neuropil regions. The supplied `refine_swc` folder contains two corresponding SWC archives, two metadata workbooks, and a source README. The NeuroXiv analysis tables and atlas files are derived or separately supplied resources; they are not part of that folder.

### Choose a skeleton version

- `refined_swc.zip` with `meta_file.xlsx`: fully refined skeletons, including multifurcation binarization. This standardizes branching topology without adding nodes or changing coordinates. Use it for general morphology, visualization, and comparable branch-order analysis. The current NeuroXiv morphology and projection tables were calculated from this version.
- `refined_swc_no_bif.zip` with `meta_file_no_bif.xlsx`: the same refinement workflow except that multifurcation binarization is omitted. Use it when retaining multifurcations or comparing topology-sensitive results matters. Branch-order results can differ between versions.

Both source workbooks contain 139,273 neuron records and a `classify` column. The non-binarized workbook omits the `Multifurcation_Binarization` processing field. The source README describes these as quality-controlled reconstructions, not error-free ground truth.

### Quality categories

- `Soma Matched`: reconstructed and annotated soma positions differ by less than 2 µm. Suitable for general morphology analyses according to the study question.
- `No Soma Annotation`: annotation coordinates are unavailable; this does not establish that the reconstructed soma is absent. Inspect before soma-dependent analysis.
- `No Soma`: manual inspection confirmed that the skeleton contains no soma. Avoid it for analyses requiring a reconstructed soma.
- `Error`: severe structural errors remain. Inspect individually and exclude where those errors would affect the analysis.

NeuroXiv retains the source `classify` value so users can choose an error-tolerance policy. The quality category and any separate processing exclusion flag answer different questions.

### SWC format and node labels

The first seven columns are node ID, type, X, Y, Z, radius, and parent ID. In the currently processed enhanced SWC archive, the file headers define type 0 as unlabeled neurite, type 1 as soma, type 2 as pre-synapse-rich, type 3 as post-synapse-rich, type 4 as mixed pre/post, type 5 as branch point, and type 6 as terminal point. Types 2–4 are custom synaptic-annotation labels, not conventional SWC axon or dendrite compartments. A node label does not represent exactly one synapse or prove signal direction.

The source README also describes a separate synaptic-polarity representation with strict pre-only, post-only, and mixed labels, and a synapse-count-augmented representation. Do not assume those representations are identical to the enhanced SWC archive without checking the file header and columns. SWC coordinates and radii used by NeuroXiv are in nanometers; divide by 1,000 to align them with the supplied micrometer-space atlas.

### NeuroXiv analysis files

`morphology_with_qc.csv` contains per-neuron morphology measurements and source quality-control fields. `projection_with_qc.csv` assigns neurite length to atlas regions; this is geometric overlap, not synaptic connectivity. `region_summary.csv` summarizes the 78 regions. Lengths and physical measurements use micrometers or micrometer-derived units as named in their columns.

The FAFA atlas archive and its mesh files are not part of this download release. The projection and region-summary CSVs are derived analysis tables; they do not contain atlas geometry. Atlas provenance and redistribution are handled separately from this release.

The original fully refined files and analysis tables retain their `files/v1/` URLs. The non-binarized pair is in `files/v2/`, with its own `checksums.sha256`; `files/v1/checksums.sha256` continues to cover the earlier release. The `processing_manifest.json` documents processing checks for the fully refined analysis tables. Source files and analysis results can also be explored in the [neuron explorer](https://drosophila.neuroxiv.org/).

### Source, license, and citations

The [supplied source README](https://drive.google.com/file/d/1UaTWAg57r4F8XL-dwUNfKo7m1WlRGkFh/view) states that the refined FlyWire v783 skeleton resource is licensed [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/). The [FlyWire public-release guidelines](https://home.flywire.ai/guidelines) also identify CC BY-NC 4.0 as the license for public-release data. Users should cite the accompanying Yuan et al. study, `Standardized FlyWire Skeletons Reveal Neuropil-Specific Organization of Synaptic Topology` (publication details pending), together with the original FlyWire publications:

- Dorkenwald et al., `Neuronal wiring diagram of an adult brain`, Nature (2024), [doi:10.1038/s41586-024-07558-y](https://doi.org/10.1038/s41586-024-07558-y).
- Schlegel et al., `Whole-brain annotation and multi-connectome cell typing of Drosophila`, Nature (2024), [doi:10.1038/s41586-024-07686-5](https://doi.org/10.1038/s41586-024-07686-5).
- For the atlas lineage: Ito et al., `A systematic nomenclature for the insect brain`, Neuron (2014), [doi:10.1016/j.neuron.2013.12.017](https://doi.org/10.1016/j.neuron.2013.12.017).

This hosted release contains both SWC refinement versions with their matching metadata workbooks, plus NeuroXiv-derived CSVs and processing documentation under CC BY-NC 4.0. The FAFA atlas archive is not included. The Yuan et al. publication details will be updated when available.
