When building or using agents in the DeepSeek Harness (DSH) plugin ecosystem, handling complex scientific workflows such as proteomics typically requires writing code or manually executing multiple steps. dsh-ezprot-plugin solves this pain point by packaging a complete proteomics analysis pipeline into a conversational interface.
What This Is¶
This is a plug-and-play plugin maintained by YukunR for DSH. It encapsulates steps such as normalization, PCA, batch correction, differential analysis, GO/KEGG enrichment, and GSEA within a conversation, allowing users to start an analysis by providing file paths and answering a few questions.
Core Features¶
- Full Workflow Encapsulation: Supports normalization → PCA → batch correction → differential analysis → GO/KEGG enrichment → GSEA.
- Automatic Environment Management: Automatically installs and manages the R 4.4.0 runtime on first use.
- Visual Tracking: Executes the pipeline step by step and provides visible summaries and charts.
- Docker Compatible: Automatically switches to Docker if local R is unavailable.
Installation and Enablement¶
Before use, ensure that Node.js (with npx), pnpm, and the DSH CLI (or DSH Desktop) are installed.
Install from the command line:
npx @deepseek-ai/dsh plugin --profile web add dsh-ezprot-plugin@0.1.1
If dsh is installed globally, you can remove the npx @deepseek-ai/ prefix:
dsh plugin --profile web add dsh-ezprot-plugin@0.1.1
After installation, restart DSH Web so that the Agent can access the proteomics_* tools.
Typical Usage¶
Users only need to provide the data file path and sample information; the Agent will run the analysis steps and write an interpretive report.
For example:
My proteomics data are in
D:\my-experiment\origin_data.txt, and the sample information is inD:\my-experiment\sample_info.txt. All samples are from mice. Compare HC and HD relative to NC.
The Agent will check and quality-control the data, confirm the comparison scheme, run the analysis step by step, and finally generate an interpretive report containing top proteins, enriched pathways, and candidate targets.
Notes¶
- Dependency Requirements: Requires Node.js, pnpm, and the dsh CLI.
- First Run: The R 4.4.0 runtime is installed on first use and takes approximately 10–20 minutes.
- Docker Option: Docker is optional; if local R cannot be configured, the plugin automatically switches to the Docker backend.
- Network-Restricted Environments: GO/KEGG background construction can be run in Docker, which is suitable for machines with restricted network access.
Summary¶
dsh-ezprot-plugin lowers the barrier to use by encapsulating the complex proteomics analysis pipeline as a conversational tool. Users do not need knowledge of R or the terminal to obtain full support from data cleaning to enrichment analysis.
