RUVprps

<!-- badges: start --> <!-- badges: end -->

RUVprps implements RUV-III with pseudo-replicates of pseudo-samples (PRPS), a novel strategy for transcriptomics data normalization when technical replicates are unavailable or poorly designed. RUVprps can also accommodate pseudo-replicate, technical replicate or a combination of them.

This user-friendly R package provides an end-to-end workflow for removing unwanted variation from large-scale transcriptomic datasets, whether derived from a single study or multiple studies. RUV-III effectively corrects for sources of variation such as library size, batch effects, and tumor purity. The package accommodates technical replicates, pseudo-replicates (PR), and pseudo-replicates of pseudo-samples (PRPS).

---

✨ Key Features

• Comprehensive diagnostic and assessment tools to evaluate both biological and unwanted variation in RNA-seq data
• Robust strategies to identify unknown sources of unwanted variation
• Unsupervised methods for identifying PRPS and negative control genes (NCGs), particularly when biological variation is unknown
• A fast and scalable implementation of RUV-III for efficient normalization of large datasets, with flexible parameter tuning
• A comprehensive numerical summary of normalization performance, helping users select the most appropriate strategy

---

📊 Overview

Overview of the functions and their applications in the RUVprps R package:

A) Tabular matrices commonly used in RNA-seq analysis, including gene expression data (assay), gene annotation, and sample annotation.
B) The function prepareSeObj() creates a SummarizedExperiment object from the tabular matrices. All outputs from RUVprps functions can be stored in the metadata of this object.
C) Functions that facilitate working with data and variables within a SummarizedExperiment object.
D) Functions for assessing unwanted variation. RUVprps provides methods tailored for both categorical and continuous variables, all accessible via the assessVariation() function.
E) Functions involved in RUV-III normalization, including identification and assessment of NCGs, PRPS construction, and the application of RUV-III. The first two steps can be performed in either supervised or unsupervised modes.

Additionally, the assessNormalization() function numerically summarizes results from the variation assessment step and ranks normalization strategies by performance. RUVprps also offers applyOtherNormalization() to apply widely used normalization methods for RNA-seq data.

---

📖 Citation

If you use RUVprps, please cite:

Trussart M, Foroutan M, Milton M, Beltran H, Speed TP, Molania R.
Fast, accurate and scalable normalization for RNA sequencing data with the RUVprps software package. bioRxiv, 2026.
https://doi.org/10.64898/2026.01.28.702384

Molania R, Foroutan M, Gagnon-Bartsch JA, Gandolfo LC, Jain A, Sinha A, Olshansky G, Dobrovic A, Papenfuss AT, Speed TP.
Removing unwanted variation from large-scale RNA sequencing data with PRPS. Nat Biotechnology. 2023;41(1):82–95.
https://doi.org/10.1038/s41587-022-01440-w

---

⚙️ Installation

After installing the required dependencies, install RUVprps from GitHub with:

library(devtools)
devtools::install_github(
  repo = "RMolania/RUVprps",
  force = TRUE,
  build_vignettes = FALSE
)