CoTRA: a comprehensive R/Shiny framework for transparent bulk and single-cell RNA-seq analysis

Abstract

Bulk RNA-seq and single-cell RNA-seq (scRNA-seq) are widely used to investigate gene-expression changes, but downstream analysis often requires multiple statistical, visualization, and reporting tools, creating fragmented workflows that are difficult to configure and reproduce. We developed CoTRA (Comprehensive Toolbox for RNA-seq Analysis), an open-source R/Shiny package providing independent bulk and scRNA-seq workflows within a common graphical environment. CoTRA supports quality control, differential expression, annotation, enrichment, dimensionality reduction, clustering, marker detection, cell-type annotation, differential abundance, trajectory inference, pathway activity, cell-cell communication, and reporting while exposing key analytical parameters. Compared with 14 other platforms across 49 predefined criteria, CoTRA fully supported 46 and partially supported three. Under matched inputs and parameters, CoTRA reproduced direct DESeq2, edgeR, and Seurat implementations, including identical significant bulk gene sets and scRNA-seq clustering (ARI = 1.000; NMI = 1.000). Retinal case studies recapitulated degeneration-associated transcriptional changes and demonstrated cell-type-resolved analysis. Synthetic scRNA-seq benchmarking scaled to 50,000 cells with 5.10 GB peak memory. CoTRA v1.0.0 requires R ≥ 4.4.0, has been tested on Linux, Windows, and macOS, is GPL-3 licensed, and is available at https://github.com/UmairSeemab/CoTRA, and support is provided through GitHub Issues. AUTHOR SUMMARY Modern sequencing technologies can measure the activity of thousands of genes across whole tissues or individual cells, but analyzing these data often requires researchers to combine many separate software tools and write substantial amounts of code. We developed CoTRA to make this process more accessible while keeping important analytical choices visible to the user. CoTRA provides graphical workflows for both bulk and single-cell RNA sequencing, covering data quality assessment, identification of expression changes, biological interpretation, cell clustering and annotation, and several advanced single-cell analyses. We tested CoTRA using published retinal datasets, compared its outputs with direct scripted analyses, and measured its computational performance as dataset size increased. The graphical workflows reproduced the corresponding scripted results when the same data and parameters were used, and the retinal examples recovered expected disease-associated expression patterns. CoTRA runs locally, so researchers do not need to upload their expression data to a mandatory external service. We hope that this combination of accessibility, transparency, and reproducible outputs will make transcriptomic analysis easier to use and inspect across collaborative biomedical research projects.