Skip to contents

Changelog

Source: NEWS.md

BadranSeq 1.0.0

Visualisation

  • do_DimPlot() provides a unified entry point for dimensionality reduction plots, routing to specialised handlers based on the reduction argument.

  • do_UmapPlot() produces publication-ready UMAP plots with cell borders, boxed cluster labels, and a clean theme out of the box.

  • do_PcaPlot() adds automatic variance-explained labels to PCA axes, formatted as “PC1 (X.X%)”.

  • do_FeaturePlot() overlays gene expression on embeddings using a viridis colour scale with cells ordered by expression.

  • EnhancedElbowPlot() extends Seurat’s elbow plot with a variance percentage axis and an optional visual cutoff line.

  • Split-panel silhouette feature: when split.by is used, each panel shows coloured cells for the current category against a grey silhouette of all cells, preserving spatial context across panels.

  • do_StatsViolinPlot() creates statistical violin plots via ggstatsplot::ggbetweenstats() with Seurat-aware data extraction, automatic colour generation, and nonparametric testing (Kruskal–Wallis + Dunn’s) by default. The group.levels argument allows comparing a subset of groups without manually subsetting the Seurat object. The pairwise.display argument controls bracket visibility.

  • do_ViolinPlot() creates descriptive violin plots with boxplot overlay, jittered raw data points, and centrality markers (median/mean diamond with numeric label). Pure ggplot2, no statistical testing.

  • do_SankeyPlot() creates Sankey (alluvial) diagrams showing the relationship between 2–3 categorical metadata variables from a Seurat object. Each stratum is labelled and coloured; flows connect cells across variables. Powered by ggalluvial.

Data extraction

  • fetch_cell_data() extracts cell-level data from Seurat objects as a tidy tibble with one row per cell. Supports optional embeddings and metadata selection. No expression data is included.

  • fetch_feature_data() extracts feature expression data as a tidy tibble with one row per cell per feature. Expression values are one column per requested layer (features long, layers wide). Supports explicit assay and layer selection, optional embeddings, and optional metadata.

  • Both functions replace Seurat::FetchData() with a cleaner interface: cell_id as a column (not rownames), tibble output, and explicit control over what data is returned.

  • get_pca_variance() returns a data.frame with variance explained and cumulative variance per principal component.

Theming

  • theme_badranseq() provides a consistent publication-ready ggplot2 theme with white background, no grids, bold titles, and legend at bottom.

  • generate_badranseq_colors() generates vivid, slightly darkened colours using the colorspace Dark 3 qualitative palette.

Interactive tools

  • seurat_sleepwalk() launches an interactive sleepwalk session for exploring cell-to-cell distances in embedding space.

  • select_cells_interactive() provides a Shiny app with brush selection for interactively selecting cells from embeddings.

Data

  • Bundled pbmc3k dataset: 3000 PBMCs preprocessed with SCTransform, PCA, UMAP, and clustered at resolution 0.5. Includes seurat_annotations with biological cell type labels.