Data from Costa et al[Macroecological invisibility in apple snails (Gastropoda: Ampullariidae): uneven occurrence and molecular coverage across public repositories]
Description
This dataset supports the manuscript “Macroecological invisibility in apple snails (Gastropoda: Ampullariidae): uneven occurrence and molecular coverage across public repositories”. The study evaluates how publicly available occurrence and molecular data represent the accepted recent diversity of Ampullariidae and how data gaps constrain macroecological inference. The dataset includes cleaned and taxonomically standardised occurrence records, a MolluscaBase-derived accepted-species reference list, GenBank and BOLD molecular records, repository/spatial-usability summaries, threshold-based visibility summaries, and sampling-completeness outputs across multiple spatial grains. The occurrence data distinguish scientific collections from observation-based records and classify records according to native, invasive or exotic distribution status. Spatial analyses are restricted to native-range scientific-collection records to avoid conflating native diversity patterns with introduced or invasive distributions. The data show that, although public repositories contain a substantial number of Ampullariidae records, coverage is highly uneven across taxa, regions and data types. The public occurrence dataset comprises 13,349 spatially unique records representing 119 accepted species, but only 3,614 native-range scientific-collection records representing 113 species were retained for the main spatial analyses. Repository presence and spatial usability are treated separately: some species are present in GBIF but lack directly usable coordinates, whereas others are not detected in the reprocessed GBIF species-level records. Literature-derived locality evidence is provided separately and was not incorporated into the spatial analyses. The molecular data also show uneven public representation. BOLD and GenBank together provide molecular records for 58 accepted species, leaving 113 species without public molecular data. The dataset should therefore be interpreted as a measure of public molecular coverage and molecular breadth, not as an estimate of true genetic diversity. BOLD BIN information is included as repository-provided barcode clustering information and should not be interpreted as formal species delimitation. Overall, the data show that Ampullariidae species become rapidly excluded from broad-scale inference as minimum occurrence thresholds increase, and that sampling completeness remains low and spatially uneven, especially in the main tropical centres of native diversity. These outputs can be used to reproduce the manuscript analyses, evaluate repository presence versus spatial usability, inspect threshold-based macroecological visibility, assess public molecular coverage, and identify priority regions and taxa for future sampling, digitisation and molecular work.
Files
Steps to reproduce
1. Download the full dataset package and keep the original folder structure. The main workflow is organised around the folder `Mendeley_Ampullariidae_Workflow/`, which contains the cleaned input datasets, hydrological masks, scripts, and output folders. 2. Place the cleaned input files inside `Mendeley_Ampullariidae_Workflow/Datasets/`. The core input files are the MolluscaBase-derived accepted-species reference list, the cleaned occurrence dataset, the curated GenBank dataset, and the curated BOLD dataset. 3. To reproduce the completeness analyses, keep the regional GLWD3 water and wetland masks inside `Mendeley_Ampullariidae_Workflow/GLWD3_water_byRegion/`. These masks are used to restrict expected richness and sampling completeness to aquatic environments. 4. Open R and set the working directory either to the parent folder containing `Mendeley_Ampullariidae_Workflow/` 5. Run the main R workflow script. The script reads the cleaned datasets, reconstructs the analytical objects, summarises repository presence and spatial usability, calculates occurrence summaries, evaluates threshold-based macroecological visibility, summarises public molecular coverage and BOLD BIN structure, and runs sampling-completeness analyses across 10, 30, 60, 110 and 220 km spatial grains. 6. For the completeness maps, keep `RUN_COMPLETENESS <- TRUE` and `SAVE_COMPLETENESS_RASTERS <- TRUE`. Saving rasters is necessary to regenerate the manuscript-ready completeness figures. Because these files can be large, temporary raster cleaning is included in the workflow to reduce disk usage during processing. 7. The script exports summary outputs to `05_final_outputs/`, manuscript tables to `06_manuscript_tables/`, and figures to `Figures/`. The main 60-km completeness figure is exported as Figure 3, while the 10-, 30-, 110- and 220-km completeness figures are exported as supplementary figures. The data-curation workflow figure was generated separately and is not produced by the R script. 8. The analyses were conducted in R using packages for data manipulation, spatial processing and plotting, including `data.table`, `dplyr`, `readxl`, `readr`, `sf`, `sp`, `raster`, `ggplot2`, `patchwork`, `scales`, `viridis` and related dependencies. The workflow does not query live GBIF, BOLD or GenBank services; it starts from the cleaned datasets deposited with this package.