Dataset DOI: 10.5061/dryad.qfttdz0v3

Description of the data and file structure

This dataset was generated for the study “Charting the nanotopography of inner hair cell synapses using MINFLUX nanoscopy”, which aimed to elucidate the three-dimensional nanoscale organisation of key synaptic proteins at inner hair cell (IHC) ribbon synapses. Using MINFLUX nanoscopy, we acquired high-precision localisations to chart the organisation of key pre- and post-synaptic proteins, including RIBEYE, RBP2, CaV1.3, Piccolino, Homer1, GluA2, and VGLUT3 to generate a molecular atlas of these specialised sensory synapses. The data was processed to extract molecular coordinates, analyse spatial clustering patterns, and derive quantitative descriptors of synaptic architecture and nanocluster organisation.

Files and variables

File: Alpha_shape_volume.xlsx

Description: Volume of alpha shape fits on coordinates of molecular counts for the proteins RIBEYE, RBP2, Homer1, CaV1.3, Piccolino and GluA2. Individual datapoints represent data from individual synapses for the respective protein analysed. 

Variables

• Protein
  Description: Name of the synaptic protein analysed (RIBEYE, RBP2, Homer1, CaV1.3, Piccolino, GluA2).
  Type: Categorical
• Alpha_Shape_Volume
  Description: Volume of the alpha shape fit generated from the 3D molecular coordinates of the given protein within a synapse.
  Unit: µm³
  Type: Numerical

File: DCLF_Test_Summary.xlsx

Description: Results of the Diggle-Cressie-Loosmore-Ford (DCLF) statistical test applied to MINFLUX data of different proteins across synapses. The DCLF test was used to assess whether molecular localisations exhibit statistically significant spatial clustering, based on deviation from complete spatial randomness over a range of distances (10 – 18 nm). Binary outcomes (TRUE/FALSE) indicate whether clustering was detected at a significance level of α = 0.002.

Variables

• File
  Description: Identifier for each synapse analysed containing protein name and numeric index.
  Type: Categorical
• DCLF True/False
  Description: Outcome of the DCLF test.
  Values:
  • TRUE: Statistically significant clustering detected
  • FALSE: No significant clustering detected
    Type: Boolean

File: List_of_epsilon_values_for_DBSCAN.xlsx

Description: List of epsilon (ε) values used for DBSCAN clustering for individual MINFLUX dataset (individuals synapses for different proteins). The epsilon value defines the neighbourhood radius used for density-based clustering and was calculated as the square root of the sum of the squares of 2σ localisation precisions in each dimension (x, y, z).

Variables

• File
  Description: Identifier for each MINFLUX dataset, combining protein name and synapse index.
  Type: Categorical
• Epsilon (nm)
  Description: Epsilon value (in nanometers) used as the neighbourhood radius parameter for DBSCAN clustering, computed from localisation precision.
  Unit: Nanometers (nm)
  Type: Numerical

File: Molecular_counts.xlsx

Description: Estimated molecular counts for each synapse imaged using 3D MINFLUX for the proteins RIBEYE, RBP2, CaV1.3, Piccolino, Homer1, and GluA2. Molecular counts were derived from processed and filtered MINFLUX localizations (see Methods) and represent the number of molecular units detected per synapse for each protein.

Variables

• Protein
  Description: Name of the synaptic protein analysed (RIBEYE, RBP2, Homer1, CaV1.3, Piccolino, GluA2).
  Type: Categorical
• Molecular Count
  Description: Number of molecular units detected in the synapse after data processing and filtering.
  Type: Integer
  Unit: Count (unitless)

File: Nanocluster_analysis.xlsx

Description: Nanocluster analysis results derived from molecular count coordinates of synaptic proteins. The analysis includes information on the number and organisation of nanoclusters, the spatial relationships between them, and intra- versus inter-cluster nearest neighbour distances. Data is presented across five worksheets.

Variables (across sheets):

1. Sheet: Number of nanoclusters
   • File: Identifier for the synapse or protein dataset
   • Number of nanoclusters: Total number of detected nanoclusters (unitless integer)
2. Sheet: Number of unclustered units
   • File: Dataset identifier
   • Unclustered units: Count of molecular units not assigned to any nanocluster (unitless integer)
3. Sheet: Units per nanocluster
   • File: Dataset identifier
   • Units per cluster: Number of molecular units per nanocluster (unitless integer)
4. Sheet: Intercluster NND
   • File: Dataset identifier
   • Intercluster NND (nm): Nearest neighbour distances between nanocluster centroids
   • Unit: Nanometers (nm)
5. Sheet: Intracluster NND
   • File: Dataset identifier
   • Intracluster NND (nm): Nearest neighbour distances between molecular units within a nanocluster
   • Unit: Nanometers (nm)

File: Nearest_neighbour_distances.xlsx

Description: Nearest neighbour distances derived from coordinates of molecular counts for the proteins RIBEYE, RBP2, Homer1, CaV1.3, Piccolino and GluA2. Each data point corresponds to a single unit and represents the Euclidean distance to its closest neighbouring unit within the same synapse.

Variables

• Protein
  Description: Name of the synaptic protein analyzed (RIBEYE, RBP2, Homer1, CaV1.3, Piccolino, GluA2).
  Type: Categorical
• NND (nm)
  Description: Nearest neighbour distance for a given molecular unit. Calculated in 3D from molecular count coordinates.
  Type: Numerical
  Unit: Nanometers (nm)

File: VGLUT3_SV_fits.xlsx

Description: Estimated synaptic vesicle (SV) diameters obtained from 2D MINFLUX localisations of VGLUT3. SV diameters were derived from circular fitting of clustered VGLUT3 localisations. Data from 4 MINFLUX images 

Variables

• VGLUT3 SV fits (vesicle diameter, nm)
  Description: Estimated diameter of individual SVs
  Type: Numerical
  Unit: Nanometers (nm)

File: 2D_2colour_VGLUT3_RIBEYE_spectral_separated.zip

Description: Contains 2D MINFLUX localisation datasets of dual-color immunolabeled synapses stained for VGLUT3 and RIBEYE post spectral separation. Individual CSV files represent individual experiments and include localisation coordinates, photon counts, and metadata (refer to DCR values to distinguish between data from respective proteins). 
Refer to the description for file 'raw_efo_cfr_filtered_data.zip' for specifications of column entries containing MINFLUX parameters. 

File: MINFLUX_Analysis_Python.zip

Description: Contains Python script to process and analyse MINFLUX localisation data. Includes routines for preprocessing, clustering (DBSCAN), alpha shape volume estimation, nearest neighbour analysis, and quantification and visualisation of nanoclusters. Also includes data from an exemplary synapse for sample run and a readme file for description of analysis routine. 

File: coordinates_units.zip

Description: Contains coordinate files with molecular localisation data for various synaptic proteins imaged using 3D MINFLUX. Each file corresponds to an individual synapse for various synaptic proteins and includes final 3D coordinates (x, y, z) of molecular counts post filtering and DBSCAN.  

File: raw_efo_cfr_filtered_data.zip

Description: Raw 3D MINFLUX output files [filtered for EFO (Effective Frequency at Offset) and CFR (Centre Frequency Ratio) from PyMINFLUX and exported]. Each CSV file represents individual synapses for each protein condition, along with metadata. 

Key for columns specifying MINFLUX variables:
tid: Trace ID, consecutive localisations belonging to one localisation burst are assigned the same Trace ID
tim: Time point of the localisation with respect to start of MINFLUX acquisition (s)
x, y, z: coordinates of localisation in three dimensions (nm)
efo: Effective Frequency at Offset (Hz) refers to the average frequency of the photons collected at the outer points of the MINFLUX pattern, corrected for background contributions, in contrast with EFC (Effective Frequency at Centre) which is measured at the center of the MINFLUX pattern.
cfr: Centre Frequency Ratio, calculated as the ratio between Effective Frequency at Centre (EFC) and Effective Frequency at Offset (EFO), i.e. EFC/EFO. Maximum CFR thresholds can be used to filter out double events and poor-quality localisations.
eco: Effective Counts at Offset are the background corrected photon counts at the MINFLUX scanning pattern except at the centre
dcr: Detector Channel Ratio, 𝐷𝐶𝑅 = (𝑐𝑜𝑢𝑛𝑡𝑠_{𝐶ℎ1,𝑜𝑓𝑓𝑠𝑒𝑡})/(𝑐𝑜𝑢𝑛𝑡𝑠_{𝐶ℎ1,𝑜𝑓𝑓𝑠𝑒𝑡} + 𝑐𝑜𝑢𝑛𝑡𝑠_{𝐶ℎ2,𝑜𝑓𝑓𝑠𝑒𝑡}). Only relevant for two-colour MINFLUX, when only one detector is in use DCR = 1.
dwell: Number of times the set dwell time in the MINFLUX sequence was repeated before the photon threshold was reached. Appropriate dwell time must be provided to pyMINFLUX upon loading a dataset.
fluo: Fluorophore identifier, only relevant for two-colour MINFLUX when using pyMINFLUX’s DCR unmixing tool.

Code/software

Raw (EFO, CFR filtered) CSV files containing 3D MINFLUX data were generated using PyMINFLUX (https://pyminflux.ethz.ch/; doi:10.5281/zenodo.7895501). 

Subsequent processing of data was performed using a custom-written Python script, which has been provided along with the submission. The ZIP file contains a Python routine for clustering (DBSCAN), alpha shape volume estimation, nearest neighbour analysis, and quantification and visualisation of nanoclusters. Data from an exemplary synapse for sample run and a readme file for description of analysis routine have also been included.

We encourage researchers to use the original MINFLUX data formats provided to ensure accurate and reproducible analyses. For support with software, workflows, or data handling, please contact the authors.