Basic Workflow | Loading the Package and Setting Up API Keys | Setting Up API Keys | Input Data Requirements | 1. Data Frame Format | 2. Seurat FindMarkers Output | 3. CSV File Path | 4. List Format | Function Parameters | Basic Usage Example | Example Output | Multi-Model Consensus Example | Consensus Output Example | Integrating with Seurat | Basic Visualization | Understanding the Output | Understanding Uncertainty Metrics | Using OpenRouter Free Models | Troubleshooting | Common Issues | Next Steps

Hierarchical Cell Type Annotation | Understanding Hierarchical Annotation | Implementing Hierarchical Annotation | Validating Hierarchical Annotations | Handling Noisy Input Data | Strategies for Noisy Marker Genes | 1. Adjust the top_gene_count parameter | 2. Apply stricter filtering for marker genes | 3. Use multi-model consensus | Handling Data with Batch Effects | 1. Use the consensus approach with a lower controversy threshold | 2. Include batch information in the tissue context | Incorporating Domain Knowledge | Using Tissue Context | Creating Custom Prompts | Combining with External Resources | Practical Case Studies | Case Study 1: PBMC Dataset Analysis | Case Study 2: Identifying Rare Cell Types | Case Study 3: Cross-Species Comparison | Performance Considerations | API Cost Management | Optimizing Runtime | Advanced Customization | Custom Processing Functions | Using the Unified Logging System | Using the CacheManager | Cache Management | Next Steps

The Multi-LLM Consensus Architecture | Why Multiple Models? | Model Diversity | The Structured Deliberation Process | 1. Initial Independent Annotation | 2. Identification of Controversial Clusters | 3. Structured Discussion for Controversial Clusters | 4. Final Consensus Formation | Uncertainty Quantification | Consensus Proportion | Shannon Entropy | Hallucination Reduction Mechanisms | Cross-Model Verification | Evidence-Based Reasoning | Critical Evaluation | Robustness to Input Noise | Collective Error Correction | Focus on Strong Signals | Uncertainty Flagging | Technical Implementation Details | Prompt Engineering | Discussion Orchestration | Comparison with Other Approaches | vs. Single LLM Annotation | vs. Traditional Annotation Methods | vs. Human Expert Annotation | Practical Implications | Next Steps

Contributing to mLLMCelltype | Getting Started | Fork and Clone the Repository | Setting Up the Development Environment | Project Structure | Development Workflow | Creating a New Feature | Code Style Guidelines | R Code Style | Documentation Guidelines | Testing | Contributing Areas | Adding Support for New LLM Models | Improving Documentation | Adding New Features | Reporting Issues | Pull Request Process | Code Review Process | Release Process | Community Guidelines | Code of Conduct | Communication Channels | Acknowledgment | License | Next Steps

General Questions | What makes mLLMCelltype different from other cell type annotation tools? | Which tissues and species does mLLMCelltype support? | How accurate is mLLMCelltype compared to other methods? | Technical Questions | How does mLLMCelltype handle cluster indices? | What is the recommended number of marker genes per cluster? | How does caching work in mLLMCelltype? | How does mLLMCelltype handle rate limits and API errors? | Performance and Optimization | How long does it take to run mLLMCelltype? | What are the API costs associated with using mLLMCelltype? | How can I use OpenRouter free models? | How can I improve the accuracy of annotations? | Troubleshooting | Why am I getting different results with the same input? | I'm getting an error about invalid cluster indices. What should I do? | How do I handle "API key not found" errors? | Why are some cell types not being correctly identified? | Integration with Other Tools | How does mLLMCelltype integrate with Seurat? | Can I use mLLMCelltype with Scanpy/AnnData in R? | How can I combine mLLMCelltype with traditional annotation methods? | Advanced Usage | How can I customize the prompts used by mLLMCelltype? | Can I add my own custom LLM models? | How can I contribute to mLLMCelltype? | Next Steps

System Requirements | Installing the R Package | Installation from CRAN (Recommended) | Installation from GitHub (Development Version) | Installation from a Local Source | Dependencies | API Keys Setup | Obtaining API Keys | Setting Up API Keys | 1. Environment Variables | 2. Direct Specification in Function Calls | 3. R Environment Variables | Verifying Installation | Common Installation Issues | Package Installation Failures | API Connection Issues | Memory Limitations | Next Steps

Overview | Installation | Setting Up API Keys | Basic Usage | Annotating Cell Types with Seurat Object | Visualizing Results | Supported Models | Advanced Usage | Using a Single LLM Model | Customizing Consensus Parameters | Using Custom Providers | Caching Results | Conclusion

Comprehensive Function Parameters | annotate_cell_types() | interactive_consensus_annotation() | Detailed Usage Scenarios | Scenario 1: Basic Annotation with a Single Model | Scenario 2: Multi-Model Consensus for High Accuracy | Scenario 2b: Using Free OpenRouter Models | Scenario 3: Working with CSV Files | Scenario 4: Custom Caching | Model Selection Guide | High Performance Models | Balanced Performance/Cost Models | Economy Models | Free Models via OpenRouter | Integration with Seurat Workflow | Advanced Parameter Tuning | Adjusting top_gene_count | Adjusting controversy_threshold | Performance Considerations | API Rate Limits and Costs | Execution Time | Troubleshooting | Common Issues with OpenRouter | Non-English Model Error Messages | Empty Results | Next Steps

Version 2.0.6 (2026-07-15) | Reliability | Maintenance | Version 1.0.0 (2023-11-15) | Initial Release | Version 1.1.0 (2024-01-20) | Features | Bug Fixes | Documentation | Version 1.2.0 (2024-03-10) | Version 1.3.0 (2024-05-15) | Version 1.4.0 (2024-07-01) | Version 1.4.1 (2024-07-15) | Breaking Changes | Version 1.2.0 | Version 1.3.0 | Version 1.4.0 | Deprecation Notices | Acknowledgments | How to Cite | Feedback and Contributions | Next Steps

Architectural Overview | Single-Agent Systems | Multi-LLM Consensus Systems | Methodological Differences | Single-Agent Approach | Consensus Approach | Performance | Cost and Resource Trade-offs | Practical Considerations | When single-agent approaches may suffice: | When consensus approaches may be preferable: | Hybrid Approaches | Summary | Next Steps

Overview | Background | Key Features | Multi-LLM Consensus Architecture | Structured Deliberation Process | Transparent Uncertainty Quantification | Other Advanced Features | Applicable Scenarios | Latest Updates | v1.1.4 (2025-04-24) | Bug Fixes | Improvements | Getting Started | Citation | Next Steps

The Challenge with Single-Model Approaches | Accuracy Limitations | Reliability Issues | The Consensus Approach: Inspired by Scientific Peer Review | The Scientific Parallel | How It Works | Why Multiple Perspectives Help | Cost Considerations | Technical Implementation | The Three-Stage Process | Quality Metrics | When to Choose Consensus | Quick Start Example | Understanding Your Results | Summary | Learn More

Introduction | Installation | One-command workflow | Stepwise pathway workflow | Convert KO abundance to KEGG pathway abundance | Run differential abundance analysis | Annotate pathway results | Visualize pathway-level results | Taxa contribution workflow | Read PICRUSt2 contribution files | Aggregate to a taxonomic level | Visualize taxa-level drivers | GSEA workflow | Summary

Introduction | Installation | Method Selection Guide | Basic GSEA Analysis (Recommended: camera method) | Covariate Adjustment | Fast Analysis with fry | Legacy: Preranked GSEA (fgsea) | Annotating GSEA Results | Visualizing GSEA Results | Pathway Label Options | Barplot | Dotplot | Enrichment Plot | Ridge Plot | Comparing GSEA and DAA Results | Conclusion | Key Recommendations | References