subsampleR

This repo provides subsampleR a flexible, modular pipeline for sub-sampling patient cohorts based on customizable criteria. This package allows users to import data from various file formats, merge it with metadata, sub-sample datasets, perform distribution statistical tests, and visualize the results.

Installation

To install the subsampleR package, you need to have R and the required libraries installed. You can directly clone the repository and use the following commands in R:

# Install dependencies if you haven't already
install.packages(c("dplyr", "ggplot2", "readr", "readxl", "arrow", "tools"))

# Load the package (assuming the `subsampleR.R` file exists in your current directory)
source("subsampleR.R")

Functions

Data Import

df <- step_import(file_path)

Imports data from various file formats (CSV, TSV, TXT, RDA, RDS, XLSX, Parquet).

Merge Data with Metadata

step_metadata(pipeline_object, metadata, cols = NULL)

Merges the input data with metadata based on the DAid column.

Sub-Sample Data

step_subsample(pipeline_object, n_samples, variable = NULL, ratio = NULL, seed = 123)

Sub-samples the data based on the specified number of samples or a ratio per category.

Perform Kolmogorov-Smirnov Test

step_ks_test(pipeline_object, population = "data", sample = "subsample_1", cols = NULL)

Performs the KS test between the population data and the sub-sampled data across specified columns.

Visualize Distributions

step_visualize(pipeline_object, population = "data", sample = "subsample_1", cols = NULL)

Visualizes the distribution of variables in both the population and sub-sampled data, generating histograms.

Case Study

To see how the subsampleR package can be utilized in practice, you can download the case study from the following link:

Case Study: SCAPIS Healthy CAC

Contact

For any questions or further information, please contact us at konstantinos.antonopoulos@scilifelab.se. Contributions are welcome, please open an issue or submit a pull request!