Changelog

Source: NEWS.md

MicrobiomeStat 1.4.1

Bug Fixes

Fixed Jensen-Shannon divergence calculation with TSS normalization

• Resolved Issue #45: NA values in Jensen-Shannon divergence after TSS normalization:
  • Fixed Kullback-Leibler divergence function to properly handle zero values in probability distributions
  • Added validation to ensure both distributions have positive values before log calculation
  • Implemented safety checks to prevent NA/infinite values in distance matrices
  • Added comprehensive test coverage for Jensen-Shannon divergence with various edge cases
  • Enhanced error handling and user warnings for invalid distance calculations

Fixed vignette execution errors

• Resolved vignette build failures in R CMD check:
  • Fixed duplicate factor level error in mStat_process_time_variable() function
  • Added proper handling of duplicate values in t0.level and ts.levels parameters
  • Improved type conversion for numeric time variables with character level specifications
  • Fixed CRAN mirror issues in vignette installation examples by adding eval=FALSE
  • Updated longitudinal analysis vignette examples to use correct parameter combinations

Additional Improvements

Complete Normalization Optimization for Barplot Functions

• Extended normalization skip to proportion data type:
  • All barplot functions (single, pair, long) now skip normalization for proportion data
  • ggplot2’s position="fill" handles proportion conversion automatically
  • Eliminates redundant TSS normalization for proportion data
  • Preserves pre-computed feature.agg.list for all data types
  • Improved user messaging for proportion data type

Consistent Behavior Across All Barplot Functions

• Unified normalization strategy:
  • count data: Skip normalization, let ggplot2 handle conversion
  • proportion data: Skip additional normalization, already in proportion format
  • other data: Skip normalization, user-processed data
  • All functions now have identical normalization handling logic

Performance and Compatibility Enhancements

• Optimal performance: Eliminated all unnecessary normalization steps in barplot functions
• Enhanced fake feature.ann compatibility: All barplot functions work seamlessly with placeholder annotations
• Data integrity: Consistent preservation of pre-computed aggregations across all functions
• Better user experience: Clear, informative messages for all data types

---

MicrobiomeStat 1.4.0

Major Improvements

Enhanced Normalization Handling in Visualization and Analysis Functions

• Optimized normalization strategy for barplot visualization:
  • Removed unnecessary TSS normalization in generate_taxa_barplot_single for count data
  • ggplot2’s position="fill" automatically handles proportion conversion for barplots
  • Preserves pre-computed feature.agg.list to avoid re-aggregation with fake feature.ann
  • Improved performance by eliminating redundant normalization steps
  • Added informative messages about normalization handling
• Fixed normalization issues in differential abundance testing:
  • Removed redundant pre-normalization in generate_taxa_test_single for count data
  • LinDA function handles normalization internally based on feature.dat.type parameter
  • Fixed hardcoded feature.dat.type = "proportion" in LinDA calls - now passes user-specified type
  • Preserves pre-computed feature.agg.list when using fake feature.ann scenarios
  • Ensures appropriate zero-value handling based on actual data type

Technical Enhancements

• Improved parameter validation:
  • Added match.arg() validation for feature.dat.type parameter
  • Better error handling and user feedback
• Enhanced compatibility with fake feature annotation scenarios:
  • Functions now work correctly when using placeholder feature.ann with pre-computed feature.agg.list
  • Prevents incorrect re-aggregation that could lead to wrong analysis results
  • Maintains data integrity across different usage patterns

Bug Fixes

• Fixed color mapping issues in barplot visualization:
  • Improved color palette assignment for better ggplot2 compatibility
  • Reduced warning messages about color scale mismatches

---

MicrobiomeStat 1.3.9

Bug Fixes

Fixed distance filtering in generate_beta_test_single

• Fixed issue where all distances in dist.obj were tested instead of only those specified in dist.name:
  • Added filtering to ensure only requested distance metrics are processed when dist.obj is provided
  • Added validation to check if all requested distances are available in dist.obj
  • Clear error messages when requested distances are not available
  • Improved code clarity with better variable naming (snake_case)
  • Fixed logical operators from & to && for proper scalar evaluation

---

MicrobiomeStat 1.3.6

Major Enhancements

Enhanced Data Aggregation Function

• Completely redesigned mStat_aggregate_data function:
  • Added meta.handle.conflict parameter with three strategies:
    • "first" (default): Use first record’s metadata, issue warnings for conflicts
    • "stop": Stop execution immediately when metadata conflicts are detected
    • "summarise": Calculate mean for numeric variables, check consistency for non-numeric
  • Fixed hardcoded subject.var limitation - now accepts any column name as subject variable
  • Preserved original feature names during aggregation (no longer modified)
  • Enhanced conflict detection with detailed reporting of which groups and variables have conflicts
  • Improved data integrity validation and error handling

Technical Improvements

• Robust conflict detection: Comprehensive metadata consistency checking across all variables
• Detailed error reporting: Clear identification of conflicting groups and variables
• Backward compatibility: All existing code continues to work with default settings
• Enhanced validation: Stricter data integrity checks prevent silent errors

User Experience

• Flexible subject variable: No longer restricted to columns named “subject”
• Informative warnings: Detailed conflict reports help users identify data quality issues
• Professional error handling: Clear, actionable error messages for data problems
• Strategy-based processing: Choose the most appropriate conflict handling for your analysis

Notes

This major update transforms mStat_aggregate_data from a basic aggregation tool into a sophisticated, configurable data processing function. The new conflict handling strategies ensure data quality and provide transparency in how metadata inconsistencies are managed, making the entire analysis pipeline more robust and trustworthy.

---

MicrobiomeStat 1.3.3

Bug Fixes and Improvements

Negative Data Handling

• Fixed negative value handling in generate_taxa_heatmap_single:
  • Added automatic detection of negative values in “other” data type
  • Disabled abundance filtering when negative values are detected
  • Updated prevalence calculation to use != 0 instead of > 0 for proper negative value handling
  • Enhanced mStat_filter function to support -Inf abundance filter for negative data
• Fixed negative value handling in generate_taxa_test_single:
  • Added automatic detection of negative values in “other” data type
  • Implemented linear model (lm) analysis for “other” data type instead of LinDA
  • Updated prevalence calculation to properly handle negative values
  • Added perform_lm_analysis helper function for robust linear model analysis

Enhanced Features

• Improved data type handling: Functions now automatically detect and appropriately handle log-transformed or other preprocessed data containing negative values
• User-friendly messaging: Added informative messages when negative values are detected and filtering is adjusted
• Backward compatibility: All changes maintain full compatibility with existing positive data workflows

Technical Improvements

• Enhanced error handling and validation for edge cases
• Improved code documentation and inline comments
• Added comprehensive test coverage for negative data scenarios

Notes

These improvements specifically address issues encountered when analyzing log-transformed microbiome data or other preprocessed data containing negative values. The package now provides robust analysis capabilities for both traditional count data and transformed abundance data.

---

MicrobiomeStat 1.3.2

Previous version features and bug fixes…

---

MicrobiomeStat 1.3.1

Previous version features and bug fixes…

---

MicrobiomeStat 1.3.0

Previous version features and bug fixes…