Frequently Asked Questions (FAQ)

Quick answers to common questions about MICOS-2024.

General Questions

What is MICOS-2024?

MICOS-2024 (Metagenomic Intelligence and Comprehensive Omics Suite) is an integrated platform for end-to-end metagenomic analysis. It combines quality control, taxonomic profiling, functional annotation, and diversity analysis into a unified workflow.

Who should use MICOS-2024?

Biologists: With minimal bioinformatics experience
Bioinformaticians: Who need standardized, reproducible workflows
Clinical researchers: Analyzing microbiome samples
Ecologists: Studying environmental microbiomes

How does MICOS-2024 compare to other tools?

Feature	MICOS-2024	Qiime2	MG-RAST	mothur
End-to-end workflow	✅	Partial	✅	Partial
WDL workflows	✅	❌	❌	❌
Docker support	✅	✅	❌	❌
CLI interface	✅	✅	Web	✅
Pipeline flexibility	High	Medium	Low	Medium

Installation Questions

What’s the recommended installation method?

Docker for:

Maximum reproducibility
Easy deployment on servers
Production environments

Conda for:

Development work
Custom modifications
Systems without Docker

How much disk space do I need?

Component	Minimum	Recommended
Software installation	5 GB	10 GB
Kraken2 database	70 GB	150 GB (PlusPF)
HUMAnN database	30 GB	50 GB
Analysis output	100 GB	500 GB
Total	~200 GB	~700 GB

Can I install without administrator privileges?

Yes! Both installation methods work without sudo:

Docker: Requires user added to docker group
Conda: Installs entirely in user directory

Do I need a GPU?

No. MICOS-2024 is CPU-based. GPU acceleration is not currently supported.

Analysis Questions

How long does analysis take?

Approximate times per sample (10M paired-end reads):

Step	Time	Notes
Quality Control	10-30 min	Depends on host genome size
Taxonomic Profiling	5-15 min	With Kraken2 standard
Functional Annotation	1-4 hours	Depends on database
Diversity Analysis	10-30 min	Per calculation
Total	2-6 hours	Highly parallelizable

Speed tips:

Use SSD for temporary files
Increase --threads parameter
Use MiniKraken for testing

What sequencing depth do I need?

Study Type	Minimum	Recommended
Pilot/testing	1M reads/sample	2M reads/sample
General profiling	5M reads/sample	10M reads/sample
Rare species detection	20M reads/sample	50M reads/sample
Functional analysis	10M reads/sample	30M reads/sample

Can I analyze single-end data?

Yes, MICOS-2024 supports both paired-end and single-end data. However:

Paired-end recommended for better taxonomic resolution
Single-end sufficient for functional profiling
Specify format in configuration or let auto-detection handle it

What about amplicon data (16S)?

While MICOS-2024 is optimized for shotgun metagenomics, it can process 16S data:

# Configuration for amplicon data
quality_control:
  kneaddata:
    bypass_trf: true    # Skip tandem repeat filter (not needed)
 
taxonomic_profiling:
  kraken2:
    confidence: 0.05    # Lower for conserved regions

For dedicated 16S analysis, consider DADA2 or QIIME2’s q2-dada2 plugin.

Database Questions

Which Kraken2 database should I use?

Database	Size	Best For
MiniKraken2 (8GB)	Small	Testing, limited RAM
Standard (70GB)	Medium	General purpose
PlusPF (100GB)	Large	Including fungi/protozoa
PlusPFP (150GB)	X-Large	Maximum coverage
Custom	Variable	Specific organisms

Can I use custom reference databases?

Yes:

# Build custom Kraken2 database
kraken2-build --download-taxonomy --db custom_db
kraken2-build --add-to-library my_genomes/*.fa --db custom_db
kraken2-build --build --threads 16 --db custom_db
 
# Use in MICOS
kraken2-db: "/path/to/custom_db"

Do I need to build databases myself?

No. Pre-built databases are available:

Kraken2: AWS Index
HUMAnN: humann_databases --download
KneadData: kneaddata_database --download

Results Questions

Why are most of my reads unclassified?

Possible reasons:

Reason	Check	Solution
Database too small	Use Standard not MiniKraken	Download larger DB
Confidence too high	Check config	Lower to 0.05-0.1
Novel organisms	Literature search	Build custom DB
Data quality	Run FastQC	Improve QC parameters
Environmental sample	Sample type	Use PlusPF database

Which diversity metric should I use?

Alpha Diversity:

Shannon: General diversity (richness + evenness)
Chao1: Richness estimation
Observed: Simple species count

Beta Diversity:

Bray-Curtis: Standard for abundance data
UniFrac: When phylogeny matters
Aitchison: For compositional data with zeros

How do I interpret statistical results?

Test	p < 0.05 means	Visualization
PERMANOVA	Groups differ significantly	PCoA plot
Kruskal-Wallis	Alpha diversity differs	Box plot
LEfSe	Specific taxa different	Bar plot

What’s a “good” diversity value?

Human gut typical ranges:

Metric	Low	Normal	High
Shannon	<2.5	2.5-4.0	>4.0
Species richness	<50	50-150	>150

Note: Context-dependent! Compare within same sample type.

Troubleshooting Questions

Analysis failed - where do I start?

Check logs: tail -n 50 logs/analysis.log
Verify inputs: Input files exist and valid
Check resources: free -h, df -h
Test with subset: Run on 1-2 samples first

Can I resume a failed analysis?

Yes, if outputs from completed steps exist:

# MICOS will skip completed steps
python -m micos.cli full-run \
  --input-dir data/raw \
  --results-dir results \
  --threads 16
  # Steps already done will be skipped

For manual control, run individual modules:

python -m micos.cli run taxonomic-profiling ...
python -m micos.cli run functional-annotation ...

How do I update MICOS-2024?

# Update code
git pull origin main
 
# Docker - rebuild image
docker compose -f deploy/docker-compose.example.yml build
 
# Conda - update environment
conda activate micos-2024
pip install -e . --upgrade

Performance Questions

Can I run MICOS on a laptop?

For small datasets (test data, <5 samples):

Minimum: 16GB RAM, 4 cores
Recommended: 32GB RAM, 8 cores

For production analysis, use servers or cloud instances.

How do I optimize for speed?

Action	Improvement	Trade-off
Use SSD for temp	2-3x faster	Need SSD storage
Increase threads	Linear to ~32	More RAM needed
Use MiniKraken	5-10x faster	Lower sensitivity
Disable gap filling	2x faster	Less pathway info

Can I run MICOS on a cluster?

Yes! Several approaches:

WDL/Cromwell: Use provided workflows
Snakemake: Convert to Snakemake workflow
Manual parallelization: Split samples, merge results

Integration Questions

Can I use MICOS output with other tools?

Yes! Output formats are standard:

Output	Format	Compatible With
Feature table	BIOM	QIIME2, phyloseq, vegan
Taxonomy	TSV	R, Python, Excel
Pathways	TSV	Gephi, Cytoscape, R

How do I import to R?

library(phyloseq)
 
# Import BIOM
physeq <- import_biom("feature-table.biom")
 
# Import metadata
sample_data <- read.csv("metadata.tsv", sep="\t", row.names=1)
sample_data(physeq) <- sample_data(sample_data)

How do I import to Python?

import pandas as pd
import biom
 
# Load BIOM table
with biom.load_table("feature-table.biom") as table:
    df = table.to_dataframe()
 
# Load taxonomy
taxonomy = pd.read_csv("taxonomy.tsv", sep="\t", index_col=0)

Citation and License

How do I cite MICOS-2024?

@software{micos2024,
  title = {MICOS-2024: Metagenomic Intelligence and Comprehensive Omics Suite},
  author = {MICOS-2024 Team},
  year = {2024},
  url = {https://github.com/BGI-MICOS/MICOS-2024},
  version = {1.1.0}
}

Also cite individual tools used (Kraken2, HUMAnN, QIIME2, etc.)

What license is MICOS-2024 under?

MIT License - free for academic and commercial use.

Still Have Questions?

Check the Installation guide
See Troubleshooting for common issues
Start a Discussion
Report an Issue

API Reference Troubleshooting