Metagenome Visualization - Usage Guide

Metagenome Visualization - Usage Guide

Overview

Visualize and statistically analyze metagenomic profiles using Python (matplotlib, seaborn, scikit-learn) or R (phyloseq, vegan, ggplot2).

Prerequisites

# Python
pip install pandas matplotlib seaborn scikit-learn scipy

# R
Rscript -e "BiocManager::install(c('phyloseq', 'microbiome'))"
Rscript -e "install.packages(c('vegan', 'ggplot2'))"

# Krona for interactive charts
conda install -c bioconda krona

Quick Start

Tell your AI agent what you want to do:

• "Create a stacked bar chart of community composition"
• "Make a heatmap of the top 20 species across samples"
• "Plot PCoA to visualize sample clustering"

Example Prompts

Composition Plots

"Create a stacked bar plot showing phylum-level composition for all samples"

"Make a heatmap of the top 15 most abundant species across my samples"

Ordination and Clustering

"Run PCoA with Bray-Curtis distance and color by treatment group"

"Cluster my samples based on species profiles and show a dendrogram"

Diversity Analysis

"Calculate and plot alpha diversity (Shannon, Simpson) for each group"

"Create rarefaction curves for all my samples"

Interactive Visualization

"Generate a Krona chart from my Kraken2 output"

"Make an interactive plot where I can hover to see species names"

What the Agent Will Do

1. Load and parse abundance data (MetaPhlAn, Bracken, or other formats)
2. Filter to relevant taxonomic levels and aggregate if needed
3. Create publication-quality visualizations
4. Save figures in requested format (PNG, PDF, SVG)

Tips

• MetaPhlAn outputs relative abundance (sums to 100%)
• Bracken outputs read counts (normalize before comparing)
• Use log transformation for highly skewed data
• phyloseq objects integrate abundance, taxonomy, and metadata
• vegan's vegdist supports many distance metrics (Bray-Curtis, Jaccard, etc.)

Common Visualizations

Resources

• phyloseq Tutorial
• vegan Documentation