The impact of package selection and versioning on single-cell RNA-seq analysis

Area of Science:

• Computational Biology
• Bioinformatics
• Genomics

Background:

• Single-cell RNA sequencing (scRNA-seq) is a powerful tool for analyzing cellular heterogeneity.
• Seurat and Scanpy are leading software packages for scRNA-seq data analysis.
• It is generally assumed that these packages produce comparable results.

Purpose of the Study:

• To compare the analytical outputs of Seurat and Scanpy across standard scRNA-seq workflow steps.
• To quantify the differences in results generated by these two widely used bioinformatics tools.
• To assess the impact of software versions on scRNA-seq analysis outcomes.

Main Methods:

• Comparative analysis of Seurat and Scanpy algorithms for scRNA-seq data processing.
• Evaluation of differences in filtering, dimensionality reduction, clustering, and differential expression analysis.
• Assessment of results variability due to software version updates.

Main Results:

• Seurat and Scanpy demonstrate substantial differences in outputs across multiple analytical steps.
• The magnitude of these differences rivals variability from reduced sequencing depth or cell number.
• Software version changes significantly impact differential expression results.

Conclusions:

• Users must critically assess bioinformatics tools used in scRNA-seq analysis.
• Prioritizing transparency and reproducibility in software development is crucial for reliable scRNA-seq research.
• The choice of scRNA-seq analysis package can significantly influence biological conclusions.