How plutzer/proximate works
ProxiMate is a GUI and command-line toolkit for analyzing proximity-labeling experiments. Users provide experimental-design and quantitative files (or SAINT-format inputs); parsing and preprocessing feed interaction scoring, annotation, and result artifacts. The GUI and MCP endpoint share a server process and dataset/session state, letting users inspect datasets, quality-control plots, enrichment, and networks; Cytoscape is an optional desktop integration. The README documents additional input formats, imputation methods, provenance, and network-analysis capabilities. Pipeline links reflect README-documented stages; unsampled implementation details are not inferred.
Main entry points: backend.py Analysis Backend · app.py Shiny App · mcp_tools.py MCP Tools · server.py Web GUI
Components 25
User Interfaces
- Web GUI
GUI/server.py - Shiny App
GUI/app.py - MCP Tools
GUI/mcp_tools.py - MCP Operations
GUI/mcp_registry.py
Analysis Pipeline
- Analysis Backend
GUI/backend.py - Input Parsing
Scripts/parse.py - Preprocessing
Scripts/protein_groups.py - Design Validation
Scripts/experimental_design.py - AFT Imputation
Scripts/refactored_aft.py - Interaction Scoring
Scripts/score.py - SAINT Model
saint/patches/SAINT-MRF-int/main.cpp - Score Annotation
Scripts/annotator.py
Results and Exploration
- QC Plots
GUI/QC_plots.py - Feature Enrichment
GUI/Ann_Enrichment.py - Network Comparison
GUI/network_comparison.py
External Integrations
- Cytoscape Controller
GUI/cytoscape_ctl.py - Cytoscape Client
GUI/cytoscape_p4c.py
Session Operations
- Dataset Store
GUI/dataset_store.py - Run Provenance
Scripts/provenance.py - Session Archive
GUI/session_archive.py - Run Artifacts
Other components
- Researcher
- MCP Agent
- Experiment Files
- Cytoscape Desktop
Connections 32
- Researcher → Web GUI: opens GUI
- MCP Agent → Web GUI: calls MCP
- Web GUI → Shiny App: starts GUI
- Web GUI → MCP Tools: registers tools
- Shiny App → Analysis Backend: uses backend
- MCP Tools → Analysis Backend: dispatches requests
- MCP Tools → MCP Operations: queries operations
- Shiny App → Dataset Store: reads and updates
- MCP Tools → Dataset Store: reads datasets
- Researcher → Experiment Files: provides
- Experiment Files → Analysis Backend: selects for analysis
- Analysis Backend → Design Validation: validates design
- Analysis Backend → Input Parsing: runs parsing
- Analysis Backend → Preprocessing: runs preprocessing
- Analysis Backend → AFT Imputation: runs imputation
- Analysis Backend → Interaction Scoring: runs scoring
- Analysis Backend → SAINT Model: runs SAINT
- Analysis Backend → Score Annotation: runs annotation
- Analysis Backend → Run Provenance: records run
- Input Parsing → Run Artifacts: writes parsed data
- Interaction Scoring → Run Artifacts: writes scores
- Score Annotation → Run Artifacts: writes annotations
- Run Provenance → Run Artifacts: writes manifest
- Shiny App → QC Plots: renders QC
- Shiny App → Feature Enrichment: renders enrichment
- Shiny App → Network Comparison: compares networks
- Shiny App → Cytoscape Controller: controls network tab
- MCP Tools → Cytoscape Controller: dispatches network tools
- Cytoscape Controller → Cytoscape Client: sends CyREST requests
- Cytoscape Client → Cytoscape Desktop: updates and queries
- Shiny App → Session Archive: manages archives
- Session Archive → Dataset Store: archives session data