Seasonal hepatic plasticity follows a descending cascade in a Neotropical frog

Data and Code for: Seasonal hepatic plasticity follows a descending cascade in a Neotropical frog Note for reviewers: Author names and institutional affiliations have been removed to comply with the double-blind review policy. They will be added upon acceptance. Overview This repository contains all raw data and reproducible R code for a manuscript submitted to Integrative Zoology. The study investigates the temporal hierarchy of seasonal phenotypic responses across four hepatic modules in the treefrog Dendropsophus minutus, using phenotypic trajectory analysis, lagged Procrustean superimposition (PROTEST), and compositional data analysis (CoDa). Repository structure . ├── README.md ├── Data_analysis_workflow.qmd # Main analysis (Quarto document) ├── lagged_protest_simulation_parallel.R # Appendix S1 — Monte Carlo validation │ ├── data/ │ ├── Biometricos.csv # Somatic indices (weight, SVL, HSI) │ ├── Melanina.csv # Melanin area per photo │ ├── Glicogenio.csv # Glycogen area per photo │ ├── Hemo.csv # Hemosiderin area per photo │ ├── Lipo.csv # Lipofuscin area per photo │ ├── Densidade_porcentagem.csv # Tissue volumetric density (point counting) │ └── Morfometria.csv # Cell and nucleus morphometry │ └── output/ # Generated figures and tables (not tracked) Data description All data refer to adult male Dendropsophus minutus collected monthly from April 2016 to March 2017 in southeastern Brazil. Collection was authorized under institutional ethics approval (details provided in the manuscript). Somatic indices (Biometricos.csv) Column Description Unit animal Individual ID — peso Body mass g tamanho Snout–vent length (SVL) mm IHS Hepatosomatic index % mes Month of collection (English) — n = 68 individuals. Histochemistry (Melanina.csv, Glicogenio.csv, Hemo.csv, Lipo.csv) Each file contains pigment or metabolite area measured from histological micrographs (5 photos per individual). Common columns: Column Description Unit animal Individual ID — fotos / foto Photo ID — mes Month of collection — estacao Season (Portuguese) — melanina / glicogenio / hemossiderina / lipofuscina Pigment area µm² area_total Total field area (denominator) µm² n = 40 individuals (subset of the 68; see manuscript for details on missing data). Tissue volumetric density (Densidade_porcentagem.csv) Relative proportions estimated by stereological point counting (Freere & Weibel 1967) on 5 histological fields per individual. Column Description Unit animal Individual ID — hepatocito Hepatocyte proportion % sinusoide Sinusoid proportion % CMM Melanomacrophage center proportion % vaso_sanguineo Blood vessel proportion % ducto Bile duct proportion % mes Month of collection — n = 68 individuals. Proportions sum to 100% per field and are treated as compositional data (CoDa) in the analysis. Cell morphometry (Morfometria.csv) Column Description Unit animal Individual ID — area_celula Hepatocyte cross-sectional area µm² area_nucleo Nucleus cross-sectional area µm² vol_celula Hepatocyte volume (estimated) µm³ vol_nucleo Nucleus volume (estimated) µm³ NCR Nucleus-to-cytoplasm ratio — mes Month of collection — n = 68 individuals. Climate data Climate variables are downloaded programmatically within the Quarto document and are not stored in this repository. Sources: Temperature, precipitation, relative humidity: BR-DWGD gridded daily data (Xavier et al. 2022), accessed via the brclimr R package. Drought index (PDSI): TerraClimate, accessed via brclimr. UV-B irradiance: glUV monthly rasters (Beckmann et al. 2014), downloaded from the UFZ data portal via terra::rast(). Photoperiod: Computed from geographic coordinates using the LightLogR R package. An active internet connection is required on first run. Subsequent runs use Quarto's chunk caching. Reproducing the analysis Requirements R ≥ 4.3.0 Quarto ≥ 1.4 R packages: install.packages(c( "tidyverse", "vegan", "smatr", "FactoMineR", "factoextra", "ecotraj", "brclimr", "LightLogR", "terra", "compositions", "zCompositions", "ggridges", "ggrepel", "gridExtra", "cowplot", "naniar", "easystats", "gt", "DiagrammeR", "DiagrammeRsvg", "rsvg", "magick" )) Running # Render the full analysis (HTML output) quarto render Data_analysis_workflow.qmd # Appendix S1: lagged PROTEST simulation (parallel version) Rscript lagged_protest_simulation_parallel.R The first render downloads climate data from external servers and may take several minutes. Cached results are reused in subsequent renders. Computational notes The main analysis completes in approximately 10–15 minutes on a modern laptop. The simulation study (Appendix S1) runs 5,000 PROTEST permutation tests (1,000 replicates × 5 scenarios) and benefits from parallelization; expect 20–40 minutes with 4+ cores. Figures generated All figures are generated from the Quarto document unless otherwise noted. Figure Source chunk Description 1 fig-data-structure Study design: 3-D data blocks 2 fig-pca-trajectories PCA trajectories for each module 3 fig-protest-network Contemporaneous PROTEST concordance network 4 fig-dag-hypothesis Hypothesized bottom-up DAG 5 fig-dag-results Observed descending-cascade DAG 6 fig-protest-lag Lagged PROTEST m² profiles 7 fig7_hierarchy (composite) Response-mode landscape (DAG + speed × lag) S1 fig-coda-dendro-density CoDa sequential binary partition S2 fig-cv-sensitivity Within-month CV sensitivity S3 fig-speed-sensitivity Trajectory speed sensitivity analyses S4 fig-phase-heatmap Phase analysis heatmap S5–S7 lagged_protest_simulation_parallel.R Appendix S1: Type I error and power License Data and code are released under the CC BY 4.0 license. If you reuse these materials, please cite the associated publication (reference to be added upon acceptance). Contact [Removed for double-blind review]