Changes in version 2.0.0                        

Major Improvements

Performance Optimization

  - Vectorized CNV classification algorithm (O(n) complexity)
  - Optimized sliding window computation with efficient matrix
    operations
  - Improved memory management with strategic garbage collection

Code Quality

  - Consistent use of TRUE/FALSE instead of T/F
  - Replaced 1:length() with seq_along() for safer iteration
  - Fixed column reference in regionToForce parameter handling

Compatibility

  - Full compatibility with Seurat 4.x and 5.x
  - Robust CreateAssayObject handling across SeuratObject versions
  - Fixed list subsetting ([[]] vs []) for proper object access

Documentation

  - Academic-style README with methodology description
  - Comprehensive vignettes with algorithm explanations
  - R-universe integration for easy installation

Bug Fixes

  - Fixed regionToForce using incorrect column name (chromosome_name →
    chromosome_num)
  - Fixed potential dimension mismatch in CNVCallingList
  - Fixed color mapping in plotCNVResultsHD for ComplexHeatmap

                        Changes in version 1.5.0                        

Bug Fixes

  - Fixed subscript out of bounds in CNVCalling
  - Fixed subscript out of bounds in CNVCallingList
  - Fixed .GetSeuratCompat() error
  - Fixed CreateAssayObject compatibility

Compatibility

  - Works with Seurat 4.4.0 + SeuratObject 4.1.4
  - Replaced SCTransform with standard workflow
  - Added comprehensive error handling

                        Changes in version 1.0.0                        

Initial Release

  - CNV inference from scRNA-seq and spatial transcriptomics data
  - Sliding window-based algorithm
  - Reference cell normalization
  - Hierarchical clustering of CNV profiles
  - Phylogenetic tree construction
  - Heatmap and tree visualization

Credits

Original fastCNV developed by Gadea Cabrejas and Clarice Groeneveld.
Enhanced and maintained by Zaoqu Liu.