Changelog

Version 0.1.0 (2025-07-03)

Initial release of MultiModulon package.

Features:

  • Multi-species/strain expression analysis using multi-view ICA

  • Bidirectional Best Hits (BBH) for ortholog detection

  • Union-Find based gene alignment across species

  • Robust multi-view ICA with clustering

  • Automated optimization of component numbers

  • Comprehensive visualization functions

  • GPU acceleration support

  • JSON-based data persistence

Core Functionality:

  • MultiModulon class for managing multi-species analysis

  • SpeciesData container for individual species data

  • BBH generation using containerized BLAST

  • Gene alignment with customizable thresholds

  • Multi-view ICA with PyTorch backend

  • Cohen’s d effect size filtering

  • Otsu’s method for threshold optimization

Visualization:

  • Gene weight plots with COG coloring

  • Activity plots with project/condition highlighting

  • Core component comparison across species

  • Batch visualization support

Optimization:

  • NRE and effect size based optimization

  • Separate optimization for core and unique components

  • Cross-validation support

  • Automated parameter selection

Utilities:

  • GFF parsing and gene table creation

  • eggNOG annotation integration

  • FASTA sequence extraction

  • Data validation and quality control

Documentation:

  • Comprehensive API documentation

  • Step-by-step tutorials

  • Example workflows