NEWS

---

SpotSweeper 1.9.3

Bug fixes

• Fixed Windows parallel nearest-neighbor search in localOutliers() and localVariance() (#27). The existing workers argument automatically selects socket workers on Windows and multicore workers on macOS/Linux; one worker continues to use serial execution. No new arguments are needed.

SpotSweeper 1.9.2

Bug fixes

• Corrected the robust local z-score calculation in localOutliers(). Scores now use the focal spot value and the unscaled median absolute deviation of neighboring spots, matching the published method. Previous versions applied the normal-consistency scaling twice and could associate a neighbor's score with the focal spot.
• Removed the spatialEco dependency, which is no longer needed for local z-score calculation.

Seurat compatibility

• Reworked Seurat support around small, tested object adapters that preserve spot names and ordering and use public SeuratObject accessors.
• Added integration tests for localOutliers(), spatial-coordinate extraction, metadata replacement, subsetting, and flagVisiumOutliers() with real Seurat objects.
• Added explicit guidance when findArtifacts() is used with data lacking mitochondrial signal, including Visium FFPE probe-panel data.

SpotSweeper 1.5.0

Major Features

• Seurat Compatibility: Added comprehensive compatibility layer enabling SpotSweeper functions to work seamlessly with Seurat spatial objects alongside existing SpatialExperiment support. This major enhancement expands SpotSweeper's usability across the spatial transcriptomics ecosystem.
• Flexible Coordinate Systems: Enhanced localOutliers() to support custom coordinate systems for neighborhood detection, enabling outlier detection in PCA, UMAP, or any reduced-dimension space.

New Functions

• getSpatialCoords(): Universal function to extract spatial coordinates from both SpatialExperiment and Seurat objects
• getMetadata(): Universal function to access metadata (colData/meta.data) from both object types
• setMetadata(): Universal function to update metadata in both object types
• subsetSpatialObject(): Universal subsetting function for both object types
• validateMetadataColumns(): Validation function to ensure required columns exist
• is_seurat() and is_spatial_experiment(): Object type detection functions

Enhanced Functions

• localOutliers():
  • Now accepts both SpatialExperiment and Seurat objects with automatic detection
  • New coords parameter allows custom coordinate matrices for neighborhood detection
  • Default behavior unchanged (uses spatial coordinates)
  • Supports PCA, UMAP, t-SNE, or any coordinate system for neighbor finding

Bug Fixes

• Fixed focal spot inclusion: Removed focal spot from neighborhood when calculating modified z-scores in localOutliers() to prevent circular logic and bias, resulting in more accurate outlier detection.

Documentation

• Updated function documentation to reflect dual compatibility and new coordinate options
• Added Seurat usage examples alongside existing SpatialExperiment examples
• Added examples for custom coordinate systems (PCA, UMAP)
• Comprehensive test suite for compatibility layer and coordinate functionality

SpotSweeper 1.3.2

New Features

• Added the 'flagVisiumOutliers()' function to identify and flag systematic outlier spots in Visium datasets. This feature enhances data quality by allowing users to efficiently detect and exclude problematic spots from downstream analyses.

Minor Changes

• Broadened Compatibility: Updated all functions to use inherits(spe, "SpatialExperiment") instead of checking class(spe) directly. This change ensures that derived classes (e.g., SpatialFeatureExperiment) are also supported, improving flexibility and ease of use.

SpotSweeper 1.3.1

Major Changes

• Function Renaming: The function plotQC has been renamed to plotQCmetrics to better reflect its purpose. The new function plotQCmetrics should be used moving forward. This change improves clarity in the package’s API by specifying that this function is designed for plotting QC metrics.

New Features and Enhancements

• shape argument: Added a shape argument to findArtifacts, allowing users to specify the neighborhood shape as either "hexagonal" or "square" for local variance calculations. This enhancement provides flexibility for different spatial arrangements in spatial transcriptomics data.

• Updated n_order parameter: Renamed the n_rings parameter to n_order in the findArtifacts function to better describe its purpose of specifying the N-order neighbors for local mitochondrial variance calculations.

• Parallelization: Added a workers argument for parallel processing using BiocParallel in both localOutlier and localVariance functions. This allows for faster computation, particularly on larger datasets.

Deprecations

• plotQC Function Deprecated: The plotQC function is now deprecated. While it remains available for backward compatibility, users are encouraged to transition to plotQCmetrics. Calling plotQC will display a warning, reminding users of the deprecation.

This change is backward compatible; existing code using plotQC will still work but will show a warning. We recommend updating your code to use plotQCmetrics to avoid any issues in future versions where plotQC may be removed.