v0.11.1

SCutils v0.11.1

🛠️ Bug fixes and improvements.

• Updated the hemoglobin and immunoglobulin (IG) gene regular expressions in CalculateQC() for more accurate human and mouse gene matching.
• Corrected the BCR and TCR gene patterns used by default in FilterVariableFeatures().

v0.11.0

SCutils v0.11.0

🛠️ Bug fixes and improvements.

• Added PlotClusterComposition() to visualize within-cluster metadata composition as stacked proportions at a selected clustering resolution.
• Improved PlotClusterComposition() input validation with explicit checks for missing cluster-resolution and grouping metadata columns.
• Updated package documentation and README function reference to include PlotClusterComposition().

SCutils v0.10.3

SCutils v0.10.3

🛠️ Bug fixes and improvements.

• Fixed a bug in CalculateQC() layer autodetection where scaled.data could be incorrectly treated as a log-normalized data layer.
• Layer selection in CalculateQC() now only accepts true log-normalized layers named like data, data.*, or data_*.
• Added stricter validation for user-provided layers in CalculateQC() to prevent invalid layer inputs from causing downstream metadata collisions.
• Prevented duplicate cell.id row-name failures in subset/scaled Seurat workflows caused by unintended inclusion of non-log layers.

v0.10.2

SCutils v0.10.1

🛠️ Bug fixes and improvements.

• FilterVariableFeatures() now does not filter out HLA genes by default. If you still want to filter out HLA genes, you can use the pattern argument to specify a custom pattern for filtering.

v0.10.1

SCutils v0.10.1

🛠️ Bug fixes and improvements.

• SelectPCs() now uses the highest integer to the estimated intrinsic dimension to account for those fractional values. It also ouputs the estimated intrinsic dimension to the console for better transparency.

v0.10.0

SCutils v0.10.0

🚀 New features.

New functions:

• VariableFeaturePlot2(): An enhanced version of Seurat's VariableFeaturePlot() that allows for labeling the top variable features on the plot. It also includes additional customization options for the plot aesthetics.

New internal functions:

• FilterVariableFeatures(): Filters the high variable features of a Seurat object after applying Seurat::FindVariableFeatures() to delete low-informative genes such as mitochondrial, ribosomal, antisense and non-coding genes.
• SelectPCs(): Estimates the intrinsic dimensionality of a Seurat object based on the PCA embeddings and returns a vector with the most informative PCs.

🛠️ Bug fixes and improvements.

• CalculateQC() now automatically uses the default assay if assay = NULL.

v0.9.1

SCutils v0.9.1

🛠️ Bug fixes and improvements.

• BatchOpenH5 now uses the future framework for parallel processing, improving performance and stability when using the multissesion mode.

v0.9.0

SCutils v0.9.0

🚀 New features.

• FeatureTest() function: This function allows users to perform feature testing on a Seurat object, enabling the identification of cells that meet specific criteria based on user-defined features and thresholds. For a given feature, users can specify a threshold and an operator (e.g., >, <, <=, ...>) to define the testing criteria. The a new column named <feature>.pass is created to indicate whether each cell passes the test. The function also supports multiple features and thresholds, allowing for complex filtering criteria. The results of the feature test are stored in the Seurat object's metadata, and users can extract and summarize these results using the ExtractFeatureTestResults() function. This function is very useful in combination with CalculateQC() to perform quality control on single-cell RNA-seq data, allows to visualize in a PCA plot or UMAP/tSNE plot the cells that pass or fail the feature test, and can be used to filter out low-quality cells from downstream analyses in a more controlled manner than directly applying filters with Seurat's subset() function.

🛠️ Bug fixes and improvements.

• ExtractFeatureTestResults() now returns the thresholds optionaly to improve compatibility with FeatureTest().
• Improved documentation of multiple functions.

v0.8.6

SCutils v0.8.6-beta

🚀 New features.

• FeatureScatterGradient() now accepts corr.method = NULL to disable correlation calculation and display. This is useful for large datasets where correlation computation may be time-consuming or unnecessary.

🏷️ Release tag

• GitHub prerelease tag/title: v0.8.6-beta.

NOTE: This is a beta release, so some features may still be under development or need further testing. We encourage users to test the new features and provide feedback. Please report any issues or bugs encountered during testing to help us improve the package.

v0.8.5-beta

SCutils v0.8.5-beta

🛠️ Bug fixes and improvements

• FeatureDensityPlot() now sorts plots alphabetically by group as FeatureScatterGradient() does, ensuring consistent ordering of plots across different visualizations.
• Now we apply coord_cartesian() instead of scale_x_continuous() and scale_y_continuous() in these functions to avoid issues with axis limits when using common.scales = TRUE. This change ensures that the axes are consistent across all plots without losing any data points or changing the density plot kernels.
• We have updated the percent.plat metric calculation in CalculateQC(). Now it calculates the percentage of counts matching the following platelet-specific genes: PPBP, PF4, GP9, ITGA2B, TUBB1, ITGB3, and GP1BA. This addition allows for better identification of platelet contamination in single-cell RNA-seq data.

🏷️ Release tag

• GitHub prerelease tag/title: v0.8.5-beta.

NOTE: This is a beta release, so some features may still be under development or need further testing. We encourage users to test the new features and provide feedback. Please report any issues or bugs encountered during testing to help us improve the package.