pondeR: Statistical Analysis Made Easier

pondeR is an R package providing a curated collection of functions for generating publication-ready descriptive and inferential statistics tables and figures — with minimal coding overhead. It is built upon the robust gtsummary framework and leverages the power of the tidyverse.

This package aims to make statistical analysis more accessible for everyone, ensuring transparency in analytical methods and providing a consistent template for results.

Key Features

pondeR offers a comprehensive suite of tools for various stages of data analysis, particularly focused on metabolomics data preprocessing and general statistical reporting:

• Direct-Injection Metabolomics Preprocessing Pipeline: Orchestrates a sequential preprocessing pipeline for metabolomics data by combining individual pondeR functions for filtering, imputation, batch correction, normalization, transformation, and scaling.
• Summarizing Data & Checking Distributions: Generate publication-ready summary tables, calculate anthropometric indices according to WHO standards, and test for normality or skewness.
• Comparing Groups & Identifying Associations: Provides methods for detecting differences between cohorts, calculating fold changes, and testing statistical associations. Includes plotting functions like volcano plots and heatmaps.
• Predictive Modeling & Classification: Offers tools for regularized regression (Ridge, LASSO, Elastic Net), Linear Mixed-Effect Models, and Logistic regression analysis with performance evaluation (AUC/AUROC) and automated bias correction options.
• Multivariate Exploration and Discrimination: Features dimensionality reduction and visualization tools (PCA, PLS, PCoA) for simplified exploration of complex datasets, including scree and score plots.
• Data Cleaning & Quality Control: Functions for filtering features based on missingness, variance, or RSD thresholds.
• Missing Value Imputation: Addresses missing data using various imputation algorithms.
• Data Normalization, Transformation, and Scaling: Transforms and scales data to ensure comparability across samples.
• Signal Drift and Batch Effects Correction: Corrects for technical variation and instrument signal drift.
• Plotting Functions: A variety of plotting utilities, including before-and-after comparisons, distribution plots, mean maps, and multi-Y plots for metabolomics and metagenomics data.
• Export Functions: Easily export data frames to Excel/CSV files and figures to PNGs.
• Utilities & Helpers: A collection of utility functions for data management, relabelling, randomization, and retrieving group sizes or significant features.

Installation

You can install the development version of pondeR directly from GitHub:

# Install devtools if you haven't already
if (!requireNamespace("devtools", quietly = TRUE)) {
  install.packages("devtools")
}

devtools::install_github("jllcalorio/pondeR")

Quick Start

Here's a quick example of how to generate a descriptive summary table using pondeR (assuming you have a data frame named my_data):

library(pondeR)
library(dplyr) # Often useful with gtsummary-based packages

# Example data (replace with your actual data)
my_data <- tibble::tibble(
  age = rnorm(100, 30, 5),
  gender = sample(c("Male", "Female"), 100, replace = TRUE),
  treatment = sample(c("A", "B", "C"), 100, replace = TRUE),
  score = runif(100, 0, 100)
)

# Generate a summary table
my_data %>%
  run_summarytable(
    # Specify variables and grouping if needed
    include = c(age, gender, score),
    by = treatment
  )

(Note: The exact arguments for run_summarytable may vary. Please refer to the package documentation for detailed usage.)

Contributing

We welcome contributions to pondeR! Please review our Code of Conduct before contributing.

Reporting Bugs

If you encounter any bugs or have feature requests, please open an issue on our GitHub Issues page.

License

This project is licensed under the MIT License.

Acknowledgements

Huge thanks to The R Project for Statistical Computing, Positron, RStudio, the developers of the gtsummary, dplyr, ggplot2, everything in the tidyverse package, and all the dependencies I fail to mention. Your work makes pondeR possible.

Future Development

More functions are continuously being added to pondeR to further enhance its capabilities and ease of use. Stay tuned for updates!