NEWS

---

scGraphVerse 0.99.21 (2025-10-21)

Bug Fixes

• Added doRNG to Imports to support GENIE3 parallel processing with nCores > 1
• Fixed igraph deprecation warning by using mode = "max" instead of mode = "undirected" for asymmetric matrices
• Fixed ggplot2 deprecation warning by replacing size with linewidth in geom_line() calls
• Added BiocStyle to Suggests for vignette building
• Added useDynLib(scGraphVerse, .registration = TRUE) to NAMESPACE for proper C code registration

Improvements

• Enhanced NAMESPACE imports to satisfy R CMD check requirements
• Improved vignette dependency management

scGraphVerse 0.99.0

🚀 Enhanced Parameter Support & Internal Dependencies

New Features

• Enhanced Parameter Support: All main functions now expose the complete parameter sets from underlying tools:
  • infer_networks(): Now supports all method-specific parameters (GENIE3, GRNBoost2, ZILGM, JRF, PCzinb)
  • create_consensus(): Now supports all INetTool parameters (tolerance, nitermax, verbose)
  • community_path(): Now supports all robin package parameters (method-specific args, comparison params, parallel processing)
• Verbose Output: Added verbose parameter to track progress across all functions
• Seed Support: Added seed parameter for reproducible results
• Internal Dependencies: Integrated ZILGM and JRF functions internally with proper GPL-2 attribution

Internal Dependencies Integration

• ZILGM Functions: Integrated find_lammax() and zilgm() functions from ZILGM package (Park et al., 2021)
• JRF Functions: Added fallback implementation for JRF functionality when package not available
• INet-Tool: Maintained as external dependency (already in imports)
• License Update: Changed to GPL (>= 2) to accommodate integrated GPL-2 code

Breaking Changes

• License changed from MIT to GPL (>= 2) due to integrated GPL-2 code
• ZILGM no longer required as external dependency (functions integrated)
• JRF now optional with internal fallback implementation

Bug Fixes & Improvements

• Improved parameter validation and error handling
• Better documentation with comprehensive parameter descriptions
• Enhanced examples showing new parameter capabilities

🚀 Initial Release

• Released the first development version of scGraphVerse.
• Provides gene regulatory network (GRN) inference from single-cell RNA-seq data.
• Supports multiple network inference methods:
  • GENIE3 (tree-based ensemble method)
  • GRNBoost2 (Python-based gradient boosting)
  • ZILGM (zero-inflated Gaussian graphical model)
  • PCzinb (partial correlation with zero-inflated negative binomial model)
  • JRF (joint random forests for multi-dataset inference)

✨ Major Features

• infer_networks(): Infer regulatory networks from count matrices.
• cutoff_adjacency(): Apply null model thresholding to weighted adjacency matrices.
• create_consensus(): Build consensus networks across methods or datasets.
• plotROC(), community_similarity(): Performance evaluation tools using ROC curves, AUC, and community structure metrics.
• plotg(), community_path(): Network visualization functions based on ggraph.

🧪 Testing and Documentation

• Added runnable examples to all major exported functions.
• Built a comprehensive README and External Dependencies installation guide.
• Set up internal unit tests for core network inference and evaluation functionalities.
• Prepared detailed documentation with reproducible examples.

🏛 Project Funding

• Supported by the National Centre for HPC, Big Data and Quantum Computing under the European Union – Next Generation EU – CN00000013 (CUP: B93C22000620006).