CASSIA v1.3.1

CASSIA v1.3.1

🚀 New Features

• Enhanced Gene ID Conversion — Improved Ensembl/Entrez ID to gene symbol conversion with auto-detection of species and Ensembl REST API fallback
• Scanpy Support — Native support for Scanpy AnnData objects and rank_genes_groups output format
• Additional Ranking Methods — Support for more differential expression ranking methods beyond FindAllMarkers
• Custom Provider Support — Configure custom OpenAI-compatible API endpoints and local LLMs (Ollama, LM Studio, etc.) in both Python and R packages

🔧 Improvements

• Symphony Compare — Enhanced Symphony mode with improved input validation and comparison workflows
• HTML Report — Improved report styling and layout for better readability
• Output Formatting — Cleaner and more structured annotation output across all agents
• Subclustering — Preserve cluster IDs in output; added key_markers and explanation columns; new additional_context parameter
• Annotation Boost — Improved defaults and auto-detection of marker data format
• Input Validation — Comprehensive validation across all agents
• Web App — Major UI/UX improvements and custom provider configuration
• Documentation — Updated with Nature Communications citation and detailed parameter tables

🐛 Bug Fixes

• Fixed cluster ID preservation in sub-clustering agent
• Fixed JSON serialization errors in batch processing
• Fixed model_settings.py path resolution issues
• Fixed subclustering output handling and formatting

📄 Citation

Published in Nature Communications: https://doi.org/10.1038/s41467-025-67084-x

Full Changelog: v1.2.0...v1.3.1

CASSIA 1.2.0

• Improved function output for clearer results.
• Better file organization for easier navigation.
• Fixed several bugs and stability issues.
• Added Smart API matching to streamline connections with different LLM APIs.

CASSIA 1.1.0

🐛 Bug Fixes
Subclustering Agent Bug Fix
Resolved an issue in the subclustering agent. The agent now performs as expected.

✨ New Features
Automatic Annotation Scoring Agent
Introduced a new agent that automatically scores annotations by comparing them with a provided gold standard. This addition helps streamline evaluation workflows and ensures consistent scoring for annotation quality checks.

CASSIA 1.0.0

1.0.0