scuttle

Provides basic utility functions for performing single-cell analyses, focusing on simple normalization, quality control and data transformations. Also provides some helper functions to assist development of other packages.

Dependencies & Environment

Package-intrinsic requirements from the Bioconductor landing page — reproduce in any R environment.

• Version: 1.22.0 · Bioconductor: 3.23 · R: ≥ 4.6
• Depends: SingleCellExperiment
• Imports: Matrix, Rcpp, BiocGenerics, S4Vectors, BiocParallel, GenomicRanges, SummarizedExperiment, S4Arrays, MatrixGenerics, SparseArray, DelayedArray, beachmat
• System requirements: C++17
• Install: BiocManager::install("scuttle")

When to Use

• Quality Control Metrics: When you need to compute basic feature-level statistics (mean counts, detection rates) using perFeatureQCMetrics or calculateAverage.
• Gene Identifier Conversion: When you need to replace stable Ensembl identifiers with unique, non-missing gene symbols using uniquifyFeatureNames.
• Data Extraction for Visualization: When you need to extract expression profiles and metadata into a standard data.frame for ggplot or model.matrix using makePerCellDF or makePerFeatureDF.
• Scaling Normalization: When removing composition biases using median-based normalization (medianSizeFactors), pooling normalization (computePooledFactors), or spike-in normalization (computeSpikeFactors).

When NOT to Use

• For state-of-the-art single-cell pipelines: Use scran instead because many of scuttle's low-level utilities are legacy functions that have been superseded by more efficient implementations in other packages.
• For complex clustering: Use quickCluster from scran instead because scuttle relies on external packages for the clustering step required prior to pooled normalization.

Data Requirements

• Input Format: A SingleCellExperiment object.
• Data Type: Raw count matrices (e.g., assay.type="counts").
• Annotations: Row names should ideally be stable identifiers (like Ensembl IDs) that can be mapped to symbols.

Key Parameters

• subsets (list(Empty=1:10)): Used in perFeatureQCMetrics to define control sets (e.g., empty wells) for calculating ratio metrics.
• assay.type ("counts"): Specifies which assay to extract data from in makePerCellDF and makePerFeatureDF.
• features ("Tspan12"): Specifies the specific genes to extract expression data for in makePerCellDF.
• cells (c(...)): Specifies the specific cells to extract expression profiles for in makePerFeatureDF.
• clusters (clusters): Used in computePooledFactors to apply normalization within predefined clusters before adjusting across clusters.

Best Practices

• Publish unfiltered count matrices with stable identifiers (e.g., Ensembl IDs) rather than filtered matrices with gene symbols, as filtered features cannot be recovered for data integration.
• Use uniquifyFeatureNames to safely convert Ensembl IDs to gene symbols for interactive analysis, ensuring duplicates are appended with identifiers and missing symbols are replaced.
• Perform a rough clustering (e.g., using scran::quickCluster) prior to running computePooledFactors on heterogeneous populations to avoid violating the assumption that most genes are non-DE.

Common Pitfalls

• Negative size factor estimates: Low-quality cells with very few expressed genes can produce negative size factors during normalization; fix this by performing strict quality control to remove damaged cells before running computePooledFactors.
• Duplicated row names: Blindly replacing stable IDs with gene symbols can create duplicate row names; fix this by wrapping the replacement in uniquifyFeatureNames.
• Violating non-DE assumptions: Running computePooledFactors globally on a highly heterogeneous dataset can distort scaling factors; fix this by supplying a clusters argument to normalize within subpopulations first.

Alternatives

• scran: Provides more advanced and actively maintained implementations of single-cell utilities, including the quickCluster function required for pooled normalization.
• DESeq2: Provides the original median ratio normalization approach that medianSizeFactors adapts for single-cell data.

Citations

• McCarthy, D. J., Campbell, K. R., Lun, A. T., & Wills, Q. F. (2017). Scater: pre-processing, quality control, normalization and visualization of single-cell RNA-seq data in R. Bioinformatics, 33(8), 1179-1186.

References

• Homepage: https://bioconductor.org/packages/scuttle
• Vignette: https://bioconductor.org/packages/release/bioc/vignettes/scuttle/inst/doc/QC.html

<!-- biomate-cta -->

---

Run this on BioMate

This skill is the knowledge layer — when, why, and how to use scuttle. To run this analysis on your own data with managed compute, automated QC, and reproducible outputs, use BioMate — free to start.

▶ Open scuttle on BioMate →