Skip to content

usage-guide.md

Latest commit

 

History

History
61 lines (38 loc) · 1.69 KB

usage-guide.md

File metadata and controls

61 lines (38 loc) · 1.69 KB

Metabolite Communication - Usage Guide

Overview

Analyze metabolite-mediated cell-cell communication using MeboCost to infer metabolic signaling between cell types from scRNA-seq data.

Prerequisites

pip install mebocost scanpy anndata

Quick Start

Tell your AI agent what you want to do:

  • "Analyze metabolite communication between my cell types"
  • "Find which metabolites are secreted by macrophages"
  • "Compare metabolic signaling between tumor and normal tissue"

Example Prompts

Basic Analysis

"Run MeboCost on my scRNA-seq data"

"Find significant metabolite-receptor interactions between cell types"

Specific Metabolites

"Which cells secrete lactate and which receive it?"

"Analyze amino acid signaling in my dataset"

Comparative Analysis

"Compare metabolite communication between treatment and control"

"Find differential metabolic signaling in tumor microenvironment"

Visualization

"Plot the metabolite communication network"

"Visualize glutamine signaling flow between cell types"

What the Agent Will Do

  1. Load and verify scRNA-seq data format
  2. Create MeboCost object with cell type annotations
  3. Run permutation-based communication inference
  4. Identify significant metabolite-receptor interactions
  5. Summarize by cell type pairs and metabolites
  6. Generate network visualizations

Tips

  • Normalization - Data must be log-normalized (scanpy standard preprocessing)
  • Cell types - Need at least 50 cells per type for robust statistics
  • Permutations - Use 1000 for publication; 100 for quick exploration
  • Gene symbols - MeboCost requires gene symbols, not Ensembl IDs
  • Species - Supports 'human' and 'mouse'