Changelog

Source: NEWS.md

LODopt 1.1.0

New features

• Added prepare_from_seurat() for converting Seurat objects directly to SummarizedExperiment input (#3).
• coef_of_interest_index and unchanged_cluster_indices now accept character names in addition to integer indices (#5).
• Added BH-adjusted p-values (adjusted_pvalue column) to results (#7).
• Optimization automatically retries with new starting values on convergence failure (#8).
• Added verbose parameter to logodds_optimized_normFactors() to control progress messages.
• Added input validation for SummarizedExperiment structure and metadata fields (#4).
• GLMM fitting is wrapped in tryCatch so a single cluster failure no longer aborts the full analysis.
• Added built-in mock_seurat_data example dataset and a Seurat workflow vignette.

Bug fixes

• Fixed operator precedence bug in simulate_cellCounts_fromTissue() that could produce incorrect tissue-level counts.
• Fixed return value and documentation mismatches (#6).

Performance

• Precompute centering matrix in estimate_variance_w_args() to avoid redundant allocation during optimization.

Infrastructure

• Added CI/CD workflow (GitHub Actions R CMD check on macOS, Windows, Ubuntu).
• Added testthat test suite with broad coverage.
• Cleaned up DESCRIPTION dependencies.

LODopt 1.0.0

• Initial release.