Loading...
Loading...
Contributors:
Dataset URI:
10.5281/zenodo.23610733OverviewThis repository contains the complete, reproducible analysis pipeline, raw datasets, and output files supporting the manuscript: "Integrating Environmental Performance Curves, Response-Surface Modelling and Machine Learning to Predict Bacterial Growth under Multiple Environmental Drivers" by Abdulsalam et al. The study reanalyses 5,230 experimental observations of growth rates across 12 freshwater heterotrophic bacterial taxa subjected to gradients of temperature (15–47 °C), pH (4–11), and salinity (0–35 g NaCl L⁻¹). The analytical workflow integrates taxon-specific quadratic performance curves, mixed-effects multidimensional response-surface modelling, and machine learning (Random Forest, XGBoost, Artificial Neural Networks) with Explainable AI (XAI) techniques. File StructureThe repository is organised as follows: Temp_Growth_Sept23.xlsx: Raw growth-rate data for the temperature gradient experiment. [From Reconciling variability in multiple stressor effects using environmental performance curves by Carmichael et al. (2025)]. pH_Growth_Sept23.xlsx: Raw growth-rate data for the pH gradient experiment. From Reconciling variability in multiple stressor effects using environmental performance curves by Carmichael et al. (2025)]. Sal_Growth_Sept23.xlsx: Raw growth-rate data for the salinity gradient experiment. From Reconciling variability in multiple stressor effects using environmental performance curves by Carmichael et al. (2025)]. 00_run_all.R: Master script. Sets the random seed (20260928) and executes the entire analysis pipeline sequentially to ensure exact reproducibility. 01_data_preparation.R: Imports, cleans, and merges the three datasets, translating experimental "Stress" labels into explicit continuous pH and salinity values, and creates leakage-proof condition_group identifiers. 02_single_stressor_models.R: Fits taxon-specific quadratic environmental performance curves and extracts optima (xoptxopt), maximum growth (μmaxμmax), and R2R2 values. 03_mixed_response_surface.R: Fits second-order and three-way mixed-effects models (using lme4/lmerTest) with taxon as a random intercept and source as a fixed effect to control for experimental batch differences. Generates 2D prediction surface plots. 04_machine_learning.R: Implements grouped cross-validation (preventing replicate leakage) and tunes Random Forest, XGBoost, and Neural Network models using the tidymodels framework. 05_explainable_ai.R: Generates model-agnostic permutation variable importance and partial dependence profiles using the DALEX and ingredients packages. outputs/: (Generated upon running the scripts) Contains all model results (.xlsx, .csv), prediction surfaces (.png, .pdf), and a session_info.txt file documenting the exact R version and package dependencies used.
Keyword(s):
MeSH Term(s):
Grants and Funding:
No Grants or Funding Data Available
Associated Publication(s):
No Associated Publication Data Available
Supplementary Data:
No Supplementary Data Available
Related Dataset(s):
No Related Dataset Data Available
The National Library of Medicine (NLM) provides information about biomedical datasets. Inclusion on an NLM Web page does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: NLM Web Policies (nih.gov)