Reproducible Data and R Code for a Meta-analysis and Machine-learning Study of Aged Micro- and Nanoplastic Phytotoxicity

Description

This dataset contains the curated analysis data and R scripts supporting a meta-analysis and machine-learning investigation of the phytotoxic effects of environmentally aged micro- and nanoplastics. The prepared dataset comprises 1,978 observations from 51 studies. The machine-learning workflow evaluates two primary effect-size targets: virgin micro- and nanoplastics relative to controls (Virgin-ES) and aged relative to virgin micro- and nanoplastics (AVV-ES). Aged-ES is retained for meta-analysis and supplementary consistency checks. The repository includes analysis scripts, execution instructions, software environment documentation, and reproducibility reports. It supports reproducibility from the prepared analysis dataset but does not reproduce the original literature search or manual extraction process.

Steps to reproduce

Download and extract the archive to a local directory. On Windows, an ASCII-only extraction path is recommended. The formal analysis input is: data/MNP_meta_analysis_dataset.csv To reproduce the meta-analysis, run in a fresh R session: Rscript --vanilla "path/to/MNP_Data_Code_Public_v1_20260922/scripts/run_meta_pipeline.R" To reproduce the machine-learning workflow, run: Rscript --vanilla "path/to/MNP_Data_Code_Public_v1_20260922/scripts/run_ml_pipeline_full.R" Required R packages and the tested software environment are documented in environment/install_packages.R and environment/VERSIONS.md. The package reproduces the analyses from the prepared observation-level dataset. It does not reproduce the original literature search, study screening, or manual extraction of values from the source publications. Detailed instructions, data definitions, version notes, and known reproducibility limitations are provided in README.md and docs/.

Institutions

• Shenzhen Technology University

  Guangdong, Shenzhen

Categories

Licence