Skip to content

usage-guide.md

Latest commit

 

History

History
81 lines (61 loc) · 2.53 KB

usage-guide.md

File metadata and controls

81 lines (61 loc) · 2.53 KB

MetaPhlAn Profiling - Usage Guide

Overview

MetaPhlAn uses clade-specific marker genes to provide accurate taxonomic profiling of metagenomic samples, outputting relative abundances that sum to 100%.

Prerequisites

conda install -c bioconda metaphlan

# Database downloads automatically on first run (~1GB)
# Or download manually:
metaphlan --install

Quick Start

Tell your AI agent what you want to do:

  • "Profile the taxonomic composition of my metagenome"
  • "Get species-level abundances for my microbial community"
  • "Merge MetaPhlAn profiles from multiple samples"

Example Prompts

Basic Profiling

"Run MetaPhlAn on sample.fastq.gz and output species abundances"

"Profile my metagenome and save the intermediate mapping file"

Multi-sample Analysis

"Process all fastq files through MetaPhlAn and merge into one table"

"Create a merged abundance table for downstream visualization"

Specific Taxonomic Levels

"Extract only species-level results from my MetaPhlAn output"

"Get genus-level abundances from the merged table"

Comparison with Kraken2

"I have both MetaPhlAn and Kraken2 results - help me compare them"

What the Agent Will Do

  1. Verify MetaPhlAn installation and database availability
  2. Run profiling with appropriate parameters for input format
  3. Generate taxonomic profile with relative abundances
  4. Merge multiple samples if requested
  5. Filter to specific taxonomic levels as needed

Tips

  • MetaPhlAn outputs relative abundance (all values sum to 100% at each level)
  • Low mapping rate is normal - only marker genes are targeted
  • Save the mapping file (--mapout) for faster re-analysis
  • UNCLASSIFIED means reads didn't match any marker gene
  • Use merge_metaphlan_tables.py to combine multiple profiles

MetaPhlAn vs Kraken2

Feature MetaPhlAn Kraken2
Method Marker genes K-mers
Output Relative abundance Read counts
Accuracy Higher at species Good overall
Database Smaller (~1GB) Larger (8-50GB)
Speed Slower Very fast

Common Issues

No Database Found

metaphlan --install

Low Mapping Rate

Normal for some samples. MetaPhlAn only considers marker genes. Check for host contamination if rate is very low.

Output All Zeros

  • Check input file is not empty
  • Verify --input_type matches file format
  • Sample may have very low microbial content

Resources