Skip to content

v0.15.4

Latest
Latest

Choose a tag to compare

View all tags
@TongZhou2017 TongZhou2017 released this 08 Feb 15:41
v0.15.4
7c60e9d

🎉 ModTector v0.15.4 - Major Update Release

We are excited to announce ModTector v0.15.4, a significant update with major feature additions, critical bug fixes, and performance improvements since v0.9.6!

🚀 Release Highlights

This release includes 6 major version updates (0.10.0 - 0.15.4) with substantial improvements:

✅ New Workflows: batch processing, single-cell support
⚡ Performance: Base quality filtering, distribution-based k-factor prediction
🔧 Format Support: Extended input format compatibility (7+ formats)
🐛 Critical Fixes: SVG metadata extraction, Siegfried method improvements
🎨 UI Enhancements: Interactive SVG with right-click menus, minimap navigation
📊 Accuracy: Improved evaluation metrics and normalization methods

🌟 Major New Features

Batch & Single-Cell Processing (v0.11.0)

  • Batch processing mode with glob pattern matching
  • Single-cell unified processing with automatic cell label extraction
  • Window-based memory management for large datasets
  • 2-3x performance improvement for single-cell data

Extended Format Support (v0.12.0, v0.12.3)

  • Native convert command with 7+ input formats:
    • RNA Framework formats (rf-rctools, rf-norm, rf-norm-xml)
    • ShapeMapper2 profile format
    • Samtools mpileup format
    • icSHAPE RT format (v0.14.7)
    • Bedgraph format (v0.14.8)
  • Automatic format detection with streaming processing
  • Dual input mode for mutation + stop signal files

Base Quality Filtering (v0.14.0)

  • Per-base quality filtering in modtector count command
  • Quality-based effective depth calculation
  • Recommended threshold: 20 (same as RNAFramework)
  • ~15% low-quality mutation filtering for improved signal-to-noise ratio

Distribution-Based K-Factor Prediction (v0.13.0)

  • Advanced k-factor prediction using statistical distribution analysis
  • Hierarchical loss function based on feature position
  • Automatic fallback to background method
  • Significantly improved accuracy, especially for 2A3 samples

Mod-Only Reactivity Workflow (v0.9.7)

  • Support for smartSHAPE datasets without unmodified controls
  • Clear command-line messaging for mod-only mode

🐛 Critical Bug Fixes

SVG Metadata Extraction (v0.15.4) ⚠️ Critical

  • Fixed base mismatch errors in SVG metadata extraction
  • Replaced context window method with individual circle tag parsing
  • Eliminates cross-contamination between adjacent circles
  • Ensures RNA_METADATA correctly matches circle data-base attributes

Siegfried Method Improvements (v0.12.2, v0.12.5, v0.14.10)

  • Negative values support: Removed forced zero-clamping, allowing negative reactivity values
  • Zarringhalam remap fix: Proper handling of negative input values
  • Integrated alignment: "align min to 0" logic integrated into zarringhalam_remap
  • Zero value percentage reduced from 54.8% to ~0.4%

Evaluation Metrics (v0.12.4)

  • Fixed incorrect sensitivity/specificity calculation
  • Now uses optimal threshold (highest F1-score) instead of fixed value > 0.0
  • Correctly handles methods where all reactivity values are positive

Stop Signal Position Correction (v0.11.1)

  • Fixed strand-specific position correction for reverse-strand reads
  • Forward strand: start_position - 1
  • Reverse strand: start_position + 1

U/T Base Handling (v0.11.3)

  • Fixed chartonum() function to handle 'U' (uracil) in RNA sequences
  • Treats U and T as equivalent, fixing 100% mutation rate issues

🎨 User Interface Improvements

Interactive SVG Enhancements (v0.14.8, v0.14.9)

  • Right-click context menu for style editing:
    • Font size and color adjustment
    • Circle radius and color modification
    • Base-specific style reset
  • Minimap navigation:
    • Draggable viewport for real-time navigation
    • Auto-hide with activity-based timer (3 seconds after idle)
  • Legend positioning: Moved to bottom-right to avoid obscuring RNA structure
  • Loading experience: Improved loading message with SVG filename display

Reactivity Distribution Preview (v0.11.4)

  • Histogram-like visualization above cutoff sliders
  • Color-coded regions based on threshold positions
  • Supports individual base and unified color range modes

📊 Accuracy & Performance Improvements

Normalization Enhancements

  • Quality-based effective depth (v0.14.4): Counts only bases with quality >= threshold
  • PCR bias correction (v0.14.4): Chi-Square distribution-based correction (optional)
  • Zarringhalam remap (v0.14.10): Integrated "align min to 0" logic for better preservation of mod/unmod differences

Evaluation Improvements

  • NaN handling (v0.12.1): Proper NaN value handling in sorting and calculations
  • Optimal threshold selection (v0.12.4): Uses F1-score optimization for all metrics
  • SNP filtering (v0.11.3): --snp-cutoff parameter to filter high mutation rate positions

🔧 Technical Improvements

Code Quality

  • Streaming processing for large files (tested up to 35GB)
  • Enhanced error handling with limited warnings strategy
  • Progress reporting for long-running operations

Workflow Integration

  • Updated Snakefile rules for new features
  • Better integration with RNA Framework, ShapeMapper2, samtools
  • Reduced external script dependencies

📦 Installation

Installation methods remain the same as v0.9.6:

# Quick Install (Recommended)
cargo install modtector

# From Source
git clone https://github.com/TongZhou2017/modtector.git
cd modtector
cargo build --release

🚀 Quick Start Examples

New: Batch Processing

modtector count \
    --batch \
    -b "samples/*.bam" \
    -f reference.fa \
    -o output_dir/ \
    -t 16

New: Format Conversion

# Convert RNA Framework output
modtector convert \
    -i input.rf \
    -o output.csv \
    -f rf-rctools \
    --rf-count-mutations

# Convert ShapeMapper2 profile
modtector convert \
    -i profile.txt \
    -o output.csv \
    -f shapemapper-profile \
    --ref-fasta reference.fa

Enhanced: Quality Filtering

modtector count \
    -b sample.bam \
    -f reference.fa \
    -o output.csv \
    --min-base-qual 20 \
    -t 8

📚 Documentation

🔄 Migration Notes

Breaking Changes

  • None - All changes are backward compatible

Recommended Updates

  • Update Snakefile rules to use new modtector convert command instead of Python scripts
  • Enable --min-base-qual 20 for improved signal quality
  • Use --k-prediction-method distribution for better k-factor prediction (especially 2A3 samples)

📈 Performance Metrics

  • Single-cell processing: 2-3x speedup over batch mode
  • Format conversion: ~1-2 million lines/second
  • Memory usage: O(1) constant for streaming operations
  • Quality filtering: ~0.9% overhead, ~15% noise reduction

🤝 Contributing

We welcome contributions! Please feel free to:

  • 🐛 Report bugs and issues
  • 💡 Suggest new features
  • 📝 Improve documentation
  • 🔧 Submit pull requests

📄 License

ModTector is licensed under the MIT License.

📞 Support

🙏 Acknowledgments

Thanks to all users who reported issues and provided feedback, especially for the SVG metadata extraction bug fixes and format conversion feature requests!

Full Changelog: https://github.com/TongZhou2017/modtector/blob/main/CHANGELOG.md

Download: See assets below for source code archives

Verify Release: SHA256 checksums available in SHA256SUMS file

Assets 2
Loading