Peripheral anatomy and central connectivity of proprioceptive sensory neurons in the Drosophila wing

Lesser, Ellen 1 ; Moussa, Anthony 1 ; Tuthill, John 1

Published Jul 03, 2025; Updated Feb 25, 2026 on Dryad. https://doi.org/10.5061/dryad.mgqnk99b5

Abstract

Recent advances in electron microscopy (EM) and automated image segmentation have produced synaptic wiring diagrams of the Drosophila central nervous system. A limitation of existing fly connectome datasets is that most sensory neurons are excised during sample preparation, creating a gap between the central and peripheral nervous systems. Here, we bridge this gap by reconstructing wing sensory axons from the Female Adult Nerve Cord (FANC) EM dataset and mapping them to peripheral sensory structures using genetic tools and light microscopy. We confirm the location and identity of known wing mechanosensory neurons and identify previously uncharacterized axons, including a novel population of putative proprioceptors that make monosynaptic connections onto wing steering motor neurons. We also find that proprioceptors of adjacent campaniform sensilla on the wing have distinct axon morphologies and postsynaptic partners, suggesting a high degree of specialization in axon pathfinding and synaptic partner matching. The peripheral location and central projections of wing sensory neurons are stereotyped across flies, allowing this wing proprioceptor atlas and genetic toolkit to guide analysis of other fly connectome datasets.

Dataset DOI: 10.5061/dryad.mgqnk99b5

Description of the data and file structure

We identified driver lines with sparse expression in the anterior dorsal mesothoracic wing nerve (ADMN) from FlyLight, a public resource showing driver line expression patterns in the central nervous system (https://www.janelia.org/project-team/flylight). We then ordered these driver lines from the Bloomington Drosophila Stock Center (BDSC), crossed them to a fluorescent reporter effector line (UAS-mCD8::GFP, or, for a few drivers, UAS-RedStinger;LexAopGFP.nls;ChAT-LexA to co-label with ChAT), and imaged the wing and wing hinge of the resulting progeny. For most images, native fluorescence was imaged. For images labelling actin, immunohistochemistry was used. See the Methods section of the accompanying manuscript (https://doi.org/10.1101/2025.05.29.656810) for details.

Files and variables

File: Zip files

Description: There is a folder for each GAL4 line containing all images collected of that line. Each z-stack is saved as a TIFF file. The filename structure for each file name is <body part imaged>. For example, 10A07_nativeGFP_wingRadius_Rfemale01_2019_03_19.tif The last channel of each filetype is autofluorescence, imaged at 647, which shows the wing itself and external structures, including bristles and campaniform sensilla. Each image is oriented such that the base of the wing is toward the left of the image and the anterior wing margin is toward the top of the image, so the tip of the wing is toward the right or the top right corner of the image.

File: driverlinetable.csv

Description: We include a table detailing which sensory structures are labeled by each driver line. Numbers in the table indicate how many neurons are labeled, e.g., 4 of 24 radius CO neurons for the first driver line, 10A07-GAL4.

Code/software

Z-Stack images can be viewed using FIJI.

File: code_Lesser_eLife_2025.zip

Description: We include a folder of analysis code with associated data files as CSV and PKL files. This resource is intended to be an archived version of the folder which is live-hosted at https://github.com/EllenLesser/Lesser_eLife_2025

The purpose of heatmap.ipynb and similarity.ipynb is to plot and analyze the data as connectivity heatmaps and similarity matrices, as in the manuscript. Each are carried out using python 3.9.6

We include each dataframe (df) as both a pkl and csv file, for ease of accessing outside of python. The pkl files were converted using the file pkl_to_csv.ipynb, included.

The data for each of these scripts is saved as in the dfs folder. Information on column titles provided below:

post_wing_table.csv / post_wing_table.pkl

id=unique identifier; created=when data was accessed; superceded_id; valid=dataset-specific code to keep track of proofreading; all data in this table has been proofread. classification system=if cell is ascending (to brain), intrinsic (to nerve cord), local (to wing neuropil), mn (motor neuron); cell_type=presumed hemilineage, not used in this manuscript; pt_supervoxel_id, pt_root_id, pt_position=dataset identifiers to find the same segment in future materialization or proofreadings.

sn_table.csv / sn_table.pkl

see above column titles, with the difference being the information contained by classification_system, here referring to the cluster based on similarity scores and cell_type=name of the peripheral sensory structure for each axon.

sn_connectivity.csv / sn_connectivity.pkl

Each row is a synaptic pair. column titles not covered above include score, which is the confidence of the predicted synapse. threshold is 15.