菜单
Advise on doublet detection and removal for scRNA (scrublet, scDblFinder, best practices).
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | doublet_detection_advice |
| description | Advise on doublet detection and removal for scRNA (scrublet, scDblFinder, best practices). |
| category | bio/single_cell_rna |
| version | 1 |
| requires_tools | ["search_knowledge_base"] |
| requires_network | false |
| user_invocable | true |
| species | any |
| modality | single_cell_rna |
| stage | qc |
| stability | experimental |
| safety_level | low |
User asks about doublets in single-cell data: how to detect, which tool, or when to remove.
Local: Use search_knowledge_base for lab SOPs on doublet detection.
Summarize: Explain that doublets are two cells in one droplet; recommend tools (Scrublet for 10x, scDblFinder for flexibility). Mention expected rate (~0.4% per 1k cells for 10x).
Workflow: Suggest: run doublet scoring → set threshold or use automatic → filter or label. Note not to over-filter.
Caveats: Brief note on multiplet rate vs sequencing depth.
Manage the BioAPEX current-feature workflow from scoping through review and completion
Turn an analysis request into a Slurm-ready execution plan with commands, resource assumptions, and job structure.
Scale a buffer recipe to a target volume and compute component masses/volumes.
Save a fetched summary or document to the knowledge base for later retrieval (e.g. after PubMed/UniProt lookup).
Interpret scRNA clusters using marker genes and suggest cell type or state.
Critically evaluate a perturbation hypothesis — challenge assumptions, propose negative controls, and flag confounders.
