single-cell-annotation-skills-with-omicverse

Guide Claude through SCSA, MetaTiME, CellVote, CellMatch, GPTAnno, and weighted KNN transfer workflows for single-cell annotation.

• Automated cell type annotation using SCSA
• Tumor microenvironment state prediction with MetaTiME
• Consensus labeling via CellVote
• Ontology mapping with CellMatch
• LLM-powered cell type annotation with GPTAnno
• Label transfer using weighted KNN

• Uses precomputed models and datasets from omicverse tutorials
• Supports multiple annotation methods including automated, consensus, and LLM-based
• Requires specific data formats and preprocessing steps
• Enables comparison of predictions with manual or known labels
• Provides visualization and interpretation tools for results
• Requires access to APIs or local models for some methods
• Must validate tissue and cell type relevance in outputs
• Offline capabilities available for some components

1. 1.Load data and perform standard preprocessing steps
2. 2.Initialize and configure the appropriate annotation model (SCSA, MetaTiME, etc.)
3. 3.Run inference to generate cell type predictions or labels
4. 4.Interpret results using visualization and comparison tools
5. 5.Validate outputs against known or expected biological labels
6. 6.Ensure correct model and data input configurations