bubble_plot

在需要基于基因表达数据画气泡图（dot plot）时使用。适用于单细胞 RNA-seq、空间转录组等场景，按细胞类型 × 癌种展示基因表达比例和均值。

Security Scan

The skill's code and runtime instructions match its stated purpose (generating bubble/dot plots from gene-expression TSV/CSV files); it requests only python3 and does not access external services or unusual credentials.

Purpose & Capability

Name/description, SKILL.md, and the included Python script are consistent: the script reads a TSV/CSV expression matrix, auto-detects metadata and numeric gene columns, and writes PNG/PDF bubble plots. Requiring python3 (and pandas/matplotlib) is proportional to the stated purpose.

Instruction Scope

SKILL.md restricts actions to checking for a local input file, ensuring Python libraries are installed, running the included script, and returning generated plot files. One operational note: the suggested pip command uses --break-system-packages (modifies system Python installation), which may be undesirable — using a virtualenv/conda environment would be safer. Otherwise the instructions do not attempt to read unrelated files or transmit data externally.

Install Mechanism

There is no install spec (instruction-only skill) and the code is included. The script depends on common Python libraries (pandas, matplotlib) and there are no downloads from external URLs or archive extraction steps in the skill package itself.

Credentials

The skill declares no environment variables, no credentials, and no config paths. The code only reads the user-supplied input file path and writes output images, which is appropriate for its function.

Persistence & Privilege

always is false and the skill does not request persistent or elevated platform privileges. It does not attempt to modify other skills or system-wide agent settings.

Assessment

This skill appears to do what it says: generate bubble/dot plots from a TSV/CSV gene-expression matrix. Before running it: (1) review the included bubble_plot.py (already present) and confirm you trust it — it only reads the provided file and writes PNG/PDF outputs; (2) avoid running the suggested pip command with --break-system-packages on a production/system Python — prefer a virtual environment (python -m venv) or conda to install pandas/matplotlib; (3) ensure the input file does not contain sensitive data you don't want processed or stored on the machine where you run the script; (4) run the skill in an isolated environment if you have any doubt about package changes or file permissions. If you need the agent to run this automatically, note it can execute the local script but has no network exfiltration code included.

Skill: 气泡图生成器 (bubble-plot)

基于基因表达矩阵数据，生成按细胞类型 × 癌种分组的气泡图（dot plot）。

何时使用

• 用户上传 TSV/CSV 表达数据文件，并要求画气泡图、dot plot、泡泡图
• 用户说"画气泡图"、"bubble plot"、"dot plot"并提供了数据文件
• 用户提供数据文件和基因列表，要求可视化表达模式

常见触发词

• 画气泡图
• bubble plot
• dot plot
• 泡泡图
• 气泡图可视化
• 表达气泡图

执行清单

1. 检查数据与环境

• 确认用户提供了数据文件（TSV/CSV）
• 确认文件已下载到工作目录
• 确认 Python 3 + pandas + matplotlib 已安装（缺失时 pip3 install --break-system-packages pandas matplotlib）

2. 了解数据结构

读取文件前几行，识别以下关键列：

角色	识别关键词	示例列名
细胞类型	majorCluster, cell_type, cluster	majorCluster
组织类型	tissue, class, condition, group	tissue
癌种	cancerType, cancer_type, disease	cancerType
基因表达	所有数值型非元数据列	FOLR1, TACSTD2, MET ...

• 如果用户指定了基因列表，只画指定的
• 如果用户未指定，自动检测所有数值型非元数据列
• 将 Adjacent 等正常组织名称统一映射为 Normal

3. 运行脚本

# 基本用法 — 自动检测所有基因
python3 skills/bubble-plot/bubble_plot.py <data.tsv>

# 只画指定基因
python3 skills/bubble-plot/bubble_plot.py <data.tsv> --genes FOLR1 TACSTD2 MET

# 自定义列名（如果自动检测失败）
python3 skills/bubble-plot/bubble_plot.py <data.tsv> --genes FOLR1 MET --cell-col CellType --cancer-col Disease

# 不按组织类型拆分
python3 skills/bubble-plot/bubble_plot.py <data.tsv> --no-split-tissue

# 自定义输出目录和分辨率
python3 skills/bubble-plot/bubble_plot.py <data.tsv> -o my_plots --dpi 600

4. 输出文件

• 默认输出到 bubble_plots/ 目录
• 每个基因 × 组织类型 = 1 个 PNG（300dpi）+ 1 个 PDF
• 文件命名：{基因}_dotplot_{Normal|Tumor}.{png|pdf}

5. 向用户报告结果

• 列出生成的文件清单和大小
• 如果有基因在数据中不存在，明确告知并跳过
• 展示 1-2 张关键图的预览

参数说明

参数	默认值	说明
input	必填	输入 TSV 或 CSV 文件路径
--genes / -g	自动检测	要画的基因名列表
--cell-col	自动检测	细胞类型列名
--cancer-col	自动检测	癌种/疾病列名
--tissue-col	自动检测	组织类型列名
--split-tissue	True	按组织类型拆分画图
--no-split-tissue	-	不拆分，所有数据画一张
--outdir / -o	bubble_plots	输出目录
--dpi	300	PNG 分辨率

图例说明

• 气泡颜色：标准化平均表达量（0-1，红=高，蓝=低，RdBu_r 色谱）
• 气泡大小：阳性细胞百分比（该细胞类型中表达量 > 0 的比例）

自检

• 数据文件是否已确认下载到本地
• 基因列名是否与数据中一致（区分大小写）
• tissue 列中 Adjacent 是否已映射为 Normal
• 输出是否包含 PNG 和 PDF 两种格式
• 缺失的基因是否已告知用户