scPairs 0.1.8 (2026-02-28)

Summary

First CRAN submission release. Addresses all reviewer feedback from the initial check, introduces a built-in test dataset, completes function documentation, and extends the test suite. Package structure has been overhauled to follow established CRAN bioinformatics conventions.

CRAN Compliance

• Added reference URL to the Description field of DESCRIPTION (<https://github.com/zhaoqing-wang/scPairs>), as required.

• Added \value tags to all three exported print() S3 methods (print.scPairs_result, print.scPairs_gene_result, print.scPairs_pair_result); each returns its input object invisibly.

• Rationalised examples: nine exported functions now run plain R examples during R CMD check. Two functions (AssessGenePair, PlotPairSummary) are wrapped in \donttest{} because their examples exceed the five-second threshold on the check servers. Three functions retain \dontrun{} because they require software not guaranteed to be present: PlotPairSpatial (spatial-assay Seurat object), PlotBridgeNetwork and PlotPairSynergy (Bioconductor annotation packages org.Hs.eg.db / org.Mm.eg.db).

• Removed set.seed() from inside .plot_bridge_network_enhanced() (R/plot_synergy.R); the package no longer modifies the user’s RNG state.

• Fixed Unicode in documentation: replaced non-ASCII characters (U+2248 approximately-equal, U+00D7 multiplication sign, U+2013/U+2014 dashes) in scpairs_testdata Rd documentation with ASCII equivalents, resolving the LaTeX PDF-manual build error.

• Fixed invalid URI in README.md: the LICENSE.md hyperlink now points to the full GitHub URL rather than a bare file path, removing the URI validity note.

• Added inst/WORDLIST listing domain-specific terms (Jaccard, KEGG, KNN, midcorrelation, transcriptomics, etc.) to suppress false-positive spell-check notes.

Built-in Test Dataset

• New dataset scpairs_testdata (data/scpairs_testdata.rda, 26 KB): a synthetic Seurat object (100 cells, 20 genes, 3 balanced clusters, 5-component PCA, 2-D UMAP) with two injected co-expression patterns:

  • GENE3 & GENE4: globally correlated across all cells (Pearson r approximately 0.89 in normalised data).
  • GENE1 & GENE2: moderately correlated within cluster 1 only.
  • All remaining pairs: near-independent noise.

• Generation script provided in data-raw/make_testdata.R (set.seed(7391)); data-raw/ is excluded from the CRAN build via .Rbuildignore.

• All runnable examples reference scpairs_testdata directly, eliminating the need for inline Seurat-object construction in any function documentation.

• Dataset documented in R/scpairs_testdata.R with @format, @details, @usage, @source, @seealso, and @family Section_0_Data tags.

Documentation Improvements

• Completed documentation for PlotPairSummary() (added @examples, improved @return) and PlotPairSynergy() (added @examples with \dontrun{}, expanded @return for all four panels).

• Added @family tags grouping all exported functions into three help families:

  • Section_0_Data – scpairs_testdata
  • Section_1_Discovery – FindAllPairs, FindGenePairs, AssessGenePair
  • Section_2_Visualization – all eleven PlotPair* and PlotBridgeNetwork functions

• Added @seealso cross-references throughout, linking each function to its most closely related counterparts.

• Improved @return descriptions for PlotPairDimplot, PlotPairViolin, PlotPairScatter, PlotPairSmoothed, PlotPairSummary, PlotPairCrossType, PlotPairSpatial, PlotPairNetwork, and PlotPairHeatmap.

Package Structure

• R/data.R renamed to R/scpairs_testdata.R, following the convention of naming dataset documentation files after the dataset itself.

• tests/README.md added: documents test categories, expected behaviour, coverage targets, and a contribution template.

• Six demonstration plots saved to docs/ from scpairs_testdata: scpairs_network.png, scpairs_heatmap.png, scpairs_dimplot.png, scpairs_smoothed.png, scpairs_violin.png, scpairs_scatter.png.

• README.md restructured: CRAN/GitHub badges, right-aligned logo, full table of contents, collapsible dependency section, Visualization Gallery embedding the six demo images, and an expanded Functions reference table.

Test Suite

• All seven test files updated to use scpairs_testdata directly, removing per-test Seurat pipeline construction and improving reproducibility. create_test_seurat() is retained in helper-test_data.R for edge-case tests requiring custom configurations (single-cluster data, spatial coordinates, unusual gene counts).

• Two latent test warnings resolved:

  • test-assess-pair.R: the n_perm = 49 call is now wrapped in suppressWarnings() – the low-permutation advisory is intentional package behaviour, not a test defect.
  • test-consolidated-functions.R: the third matrix row in the .compute_ratio_consistency batch test now uses abs(rnorm(n)) to prevent log2(mat + 1) from receiving negative inputs.

• 115 tests passing, 0 failures, 0 warnings, 0 skips.

scPairs 0.1.7 (2026-02-23)

Bridge Gene Network Visualization

• New exported function PlotBridgeNetwork(): standalone bridge gene network for a focal gene pair. Nodes are positioned via MDS on Jaccard pathway distance; radial depth reflects shared-term count; focal genes appear at the centre. Bridge gene nodes are sized by bridging strength and filled by mean expression (viridis palette). Focal-to-bridge edges are coloured by source gene (red/blue) and scaled in width by shared term count; dotted inter-bridge edges encode pairwise Jaccard similarity. Supports a layout parameter and configurable sim_threshold, pt_size_range, and edge_width_range.

• Refactored .plot_bridge_network_enhanced(): replaces the prior static circular layout with MDS-radial positioning; adds weighted edges, expression-coded fill, duplicate-edge deduplication, and bridge-bridge similarity overlay.

• PlotPairSynergy() panel 2 now calls .plot_bridge_network_enhanced().

scPairs 0.1.6 (2026-02-11)

Unified Computation Engine & API Consolidation

• Shared metric engine (compute_metrics.R): all three discovery functions now share a single .compute_pair_metrics() backend, eliminating approximately 600 lines of duplicated code.

• mode parameter added to all discovery functions: "all" (default), "expression", "prior_only".

• Unified output format: AssessGenePair() now returns a $pairs data.table with the same column schema as FindAllPairs() and FindGenePairs(). All visualisation functions accept any of the three result classes interchangeably.

• Removed pheatmap from Suggests (unused dependency).

• Internal helpers added: .resolve_cluster_ids(), .extract_pair_vals(), .extract_pairs_df().

scPairs 0.1.5 (2026-02-11)

Prior Knowledge Integration & Synergy-Aware Scoring

• Prior knowledge layer: prior_score (GO/KEGG Jaccard co-annotation, weight 2.0) and bridge_score (pathway bridge genes, weight 1.8).

• New metric neighbourhood_synergy (weight 1.5).

• New visualisation PlotPairSynergy(): 6-panel evidence dashboard.

• Package now integrates 14 metrics across 5 evidence layers.

• use_prior, organism, and custom_pairs parameters added to all three main discovery functions.

scPairs 0.1.4 (2026-02-08)

Robustness & Input Validation

• min_pct_expressed filter added to exclude sparse genes from analysis.
• Comprehensive input validation with informative error messages.
• Centralised schema (schema.R) and shared plot utilities (plot_utils.R).

scPairs 0.1.3 (2026-02-08)

Cross-Cell-Type Interaction Metric

• New metric cross_celltype_score: measures trans-cellular synergy between cell populations (weight 1.5).
• New visualisation PlotPairCrossType(): cross-cell-type interaction heatmap.

scPairs 0.1.2 (2026-02-08)

Neighbourhood-Aware Metrics

• Three new metrics: smoothed_cor, neighbourhood_score, cluster_cor.
• Two new visualisations: PlotPairSmoothed(), PlotPairSummary().

scPairs 0.1.1 (2026-02-08)

Performance Optimizations

• 5–20x speedups achieved through vectorisation of all core metric computations.

scPairs 0.1.0 (2026-02-07)

Initial Release

• Three core workflows: FindAllPairs(), FindGenePairs(), AssessGenePair().
• Seven metrics: Pearson, Spearman, biweight midcorrelation, mutual information, ratio consistency, Lee’s L, and co-location quotient.
• Six visualisation functions.