NEWS
PopComm 1.0.1
Bug fixes filter_lr_single(), score_lr_single(), one_step_single()
- Fixed duplicated samples in projection scores caused by overlapping sender or receiver cell type names.
- Replaced substring-based column matching with exact suffix matching for sender and receiver average expression.
- Added explicit checks to stop execution when duplicated sample names are detected.
Changes in heatmap_sample(), boxplot_lr_group_comparison()
heatmap_sample(): Added an automatic clustering switch. When the number of rows or columns is less than 2, the corresponding cluster_rows or cluster_cols is automatically set to FALSE and the dendrogram height is set to 0, preventing hclust errors.
boxplot_lr_group_comparison(): Added two new parameters:
show_counts — displays sample counts (n=) for each group (default TRUE).
stat_label — allows flexible significance labeling styles: "p.signif" (asterisks), "p.format" (formatted p-value), "p.value" (raw p-value), or "none" (no significance label).
PopComm 1.0.0 (2025-11-01)
Changes in filter_lr_single(), filter_lr_all(), score_lr_single(), single_lr_all(), one_step_single(), one_step_all()
- Added support for average expression matrices. These functions now accept either Seurat objects or average expression matrices (a numeric matrix of gene expression where columns encode cell types and samples, and rows are genes).
- No changes to existing Seurat-based workflows.
PopComm 0.1.2
Changes in filter_lr_single(), filter_lr_all(), one_step_single(), one_step_all()
- Added new parameters:
min_r2 and min_fstat to filter models by minimum R-squared and F-statistic thresholds.
- Output now includes r_squared (
r2) and f_statistic (fstat) for each linear regression model.
- Models with R² below
min_r2 or F-statistic below min_fstat are excluded from results.
- Enhances model screening based on goodness-of-fit and overall regression significance.
PopComm 0.1.1
Changes in circle_plot()
- Updated
circle_plot() to support compatibility with igraph version 2.1.5 and above.
- For
igraph <= 2.1.4, self-loop angles are calculated manually to radiate outward based on node positions.
- For
igraph >= 2.1.5, loop angles are handled automatically by igraph; added logic to suppress or delete self-loops when disabled.
- Added internal version check using
utils::packageVersion() to differentiate logic paths and ensure backward compatibility.
- Removed direct assignment to
E(g)$loop.angle for new igraph versions, which now throw errors with incorrect defaults.
- Added
@importFrom utils packageVersion directive to support version checking.
- Improved robustness of layout-based self-loop rendering.