Day 8: Workflow Wrappers & Web Platforms

📚 Table of Contents

• Introduction
• Part I: Workflow Wrappers
• Part II: Web-Based Platforms
• Part III: Choosing the Right Tool
• Best Practices

🎯 Introduction

You’ve learned the individual tools. Now let’s automate everything.

The Challenge

Days 1-7 taught you:

• Quality control
• Assembly
• Binning
• Annotation
• Comparative analysis

But running each tool separately is:

• ❌ Time-consuming
• ❌ Error-prone
• ❌ Hard to reproduce
• ❌ Requires deep Linux knowledge

The Solution

Workflow wrappers = Pre-built pipelines that automate everything
Web platforms = Run analyses in your browser (no installation!)

🔧 Part I: Workflow Wrappers

Overview

🎁 1. MetaWRAP - Complete Metagenome Pipeline

MetaWRAP wraps Days 1-7 into simple commands.

Installation

# Create conda environment
conda create -n metawrap-env python=2.7
conda activate metawrap-env

# Install MetaWRAP
conda install -c ursky metawrap-mg

# Download databases (WARNING: ~300 GB!)
metaWRAP_db_setup.sh

Complete Workflow

please use metaWrap_complete_workflow.sh from Day 8 →

MetaWRAP Modules

Available modules:

# Quality control
metawrap read_qc

# Assembly
metawrap assembly --metaspades  # or --megahit

# Binning
metawrap binning --metabat2 --maxbin2 --concoct

# Refinement
metawrap bin_refinement -c 50 -x 10

# Reassembly
metawrap reassemble_bins

# Quantification
metawrap quant_bins

# Classification
metawrap classify_bins

# Annotation
metawrap annotate_bins

# Bin visualization
metawrap blobology

Advantages

✅ All-in-one solution
✅ Integrates multiple tools
✅ Automatic bin refinement
✅ Good documentation

Disadvantages

❌ Large database requirements (~300 GB)
❌ Python 2.7 dependency
❌ Can be slow
❌ Less flexible than individual tools

🔁 2. nf-core/mag - Nextflow Pipeline

Nextflow = Modern workflow manager (reproducible, scalable)

Installation

# Install Nextflow
curl -s https://get.nextflow.io | bash
mv nextflow ~/bin/

# Or via conda
conda install -c bioconda nextflow

Running nf-core/mag

# Basic run
nextflow run nf-core/mag \
    --input samplesheet.csv \
    --outdir results \
    -profile docker

# Full options
nextflow run nf-core/mag \
    --input samplesheet.csv \
    --outdir results \
    --skip_spades \
    --megahit_fix_cpu_1 \
    --min_contig_size 1500 \
    --min_length_unbinned_contigs 1000000 \
    --max_unbinned_contigs 100 \
    --busco_db bacteria_odb10 \
    -profile docker,test

Sample Sheet Format

sample,group,short_reads_1,short_reads_2,long_reads
sample1,group1,s1_R1.fastq.gz,s1_R2.fastq.gz,
sample2,group1,s2_R1.fastq.gz,s2_R2.fastq.gz,s2_long.fastq.gz
sample3,group2,s3_R1.fastq.gz,s3_R2.fastq.gz,

nf-core/mag Features

✅ Modern workflow engine
✅ Automatic resume (restart from failures)
✅ Multiple assemblers (MEGAHIT, SPAdes)
✅ Multiple binners (MetaBAT2, MaxBin2, CONCOCT)
✅ BUSCO quality checks
✅ GTDB-Tk classification
✅ MultiQC reports

Output Structure

results/
├── QC/
│   ├── fastp/
│   └── fastqc/
├── Assembly/
│   ├── MEGAHIT/
│   └── SPAdes/
├── Binning/
│   ├── MetaBAT2/
│   ├── MaxBin2/
│   └── bins_summary.tsv
├── GenomeBinning/
│   ├── bins/
│   └── checkm2_results/
├── Taxonomy/
│   └── GTDB-Tk/
└── MultiQC/
    └── multiqc_report.html

🖥️ 3. Anvi’o - Interactive Metagenomics

Anvi’o = Analysis + Visualization platform

Installation

conda create -y -n anvio-8 python=3.10
conda activate anvio-8
conda install -y -c conda-forge -c bioconda anvio=8
anvi-self-test

Quick Workflow

# 1. Create contigs database
anvi-gen-contigs-database -f contigs.fa -o contigs.db

# 2. Annotate
anvi-run-hmms -c contigs.db -T 8
anvi-run-ncbi-cogs -c contigs.db -T 8
anvi-run-kegg-kofams -c contigs.db -T 8

# 3. Map reads
bowtie2-build contigs.fa contigs
bowtie2 -x contigs -1 R1.fastq -2 R2.fastq -S sample.sam -p 8
samtools view -bS sample.sam | samtools sort -o sample.bam
anvi-init-bam sample.bam -o sample_sorted.bam

# 4. Profile
anvi-profile -i sample_sorted.bam -c contigs.db -o PROFILE

# 5. Binning (automatic)
anvi-cluster-contigs -p PROFILE/PROFILE.db -c contigs.db \
    -C CONCOCT --driver CONCOCT -T 8

# 6. Interactive visualization
anvi-interactive -p PROFILE/PROFILE.db -c contigs.db

Anvi’o for Pangenomics

# Create genomes storage
anvi-gen-genomes-storage -e external-genomes.txt -o GENOMES.db

# Run pangenome
anvi-pan-genome -g GENOMES.db \
    --project-name my_pangenome \
    --num-threads 8

# Visualize
anvi-display-pan -g GENOMES.db -p my_pangenome/my_pangenome-PAN.db

Anvi’o Strengths

✅ Beautiful interactive interface
✅ Real-time binning (click to bin!)
✅ Metabolic reconstruction
✅ Pangenomics integration
✅ Publication-quality figures

🐍 4. ATLAS - Scalable Metagenomics

ATLAS uses Snakemake for large-scale analyses.

Installation

conda create -n atlas -c bioconda -c conda-forge metagenome-atlas
conda activate atlas
atlas init --db-dir databases
atlas download --db-dir databases

Running ATLAS

# Initialize project
atlas init --db-dir ~/databases my_project

# Edit config.yaml and samples.tsv

# Run complete workflow
atlas run all \
    --working-dir my_project \
    --config-file my_project/config.yaml \
    --cores 32 \
    --jobs 5

# Or specific steps
atlas run qc            # Quality control
atlas run assembly      # Assembly
atlas run binning       # Binning
atlas run genomes       # Genome refinement

ATLAS Features

✅ Automatic database downloads
✅ Co-assembly support
✅ Metatranscriptomics (RNA-seq integration)
✅ DRAM annotation
✅ Easy configuration

🧩 5. SqueezeMeta - Integrated Pipeline

SqueezeMeta = Complete pipeline with co-assembly.

Installation

git clone https://github.com/jtamames/SqueezeMeta.git
cd SqueezeMeta
./install_linux.sh

Running SqueezeMeta

# Single sample
SqueezeMeta.pl \
    -m sequential \
    -p project_name \
    -s samples.txt \
    -f raw_reads \
    -t 24

# Co-assembly mode
SqueezeMeta.pl \
    -m coassembly \
    -p coassembly_project \
    -s samples.txt \
    -f raw_reads \
    -t 24 \
    --assembly megahit

SqueezeMeta Outputs

• Taxonomic classification (multiple databases)
• Functional annotation (KEGG, COG, Pfam)
• Metabolic pathways
• Interactive HTML reports
• Abundance tables

🌐 Part II: Web-Based Platforms

Why Use Web Platforms?

✅ No installation (works in browser)
✅ Pre-configured (databases ready)
✅ User-friendly (GUI interface)
✅ Collaborative (share projects easily)
✅ HPC access (use their servers)

❌ Limited customization
❌ Data upload time (can be slow)
❌ Storage limits (usually)
❌ Privacy concerns (your data on their servers)

🌌 1. Galaxy - Web-Based Analysis

Galaxy = Most popular web platform for bioinformatics

Getting Started

Public servers:

• Main: https://usegalaxy.org
• Europe: https://usegalaxy.eu
• Australia: https://usegalaxy.org.au

Create account → Free!

Galaxy Metagenomics Workflow

Step 1: Upload Data

1. Click “Upload Data”
2. Choose files or paste URLs
3. Select “fastqsanger.gz” format
4. Click “Start”

Step 2: Quality Control

Tool: FastQC

Tools → FASTQ Quality Control → FastQC
- Input: Your FASTQ files
- Execute

Tool: Trim Galore!

Tools → FASTQ Quality Control → Trim Galore!
- Single-end or Paired-end: Paired-end
- Reads 1: R1.fastq
- Reads 2: R2.fastq
- Execute

Step 3: Assembly

Tool: MEGAHIT

Tools → Assembly → MEGAHIT
- Single/Paired-end: Paired
- Forward: trimmed_R1.fastq
- Reverse: trimmed_R2.fastq
- Minimum contig length: 1500
- Execute

Step 4: Binning

Tool: MetaBAT2

Tools → Metagenomics → MetaBAT2
- Assembly: contigs.fasta
- BAM files: mapped_reads.bam
- Minimum contig length: 2500
- Execute

Step 5: Quality Check

Tool: CheckM

Tools → Metagenomics → CheckM lineage_wf
- Bins: MetaBAT2 bins
- Execute

Step 6: Annotation

Tool: Prokka

Tools → Annotation → Prokka
- For each bin
- Kingdom: Bacteria
- Execute

Galaxy Workflows

Create reusable workflows:

1. Go to Workflow menu
2. Click “Extract Workflow”
3. Select history
4. Name workflow
5. Save

Run on new data:

1. Upload new files
2. Go to Workflows
3. Select your workflow
4. Map inputs
5. Run!

Galaxy Advantages

✅ No installation
✅ 1000+ tools available
✅ Workflows can be shared
✅ Educational tutorials
✅ Free compute resources

🧬 2. KBase - Systems Biology Platform

KBase = DOE Joint Genome Institute platform

Website: https://kbase.us

Getting Started

1. Create account (free)
2. Create narrative (like Jupyter notebook)
3. Upload data
4. Add analysis apps

KBase Metagenomics Apps

Available Apps:

1. Trim Reads (Trimmomatic)
2. Assemble Reads (MEGAHIT, SPAdes)
3. Annotate Assembly (RAST, PROKKA)
4. Bin Contigs (MaxBin2, MetaBAT2)
5. Classify Taxonomy (GTDB-Tk)
6. Compare Genomes (Pangenome)
7. Build Metabolic Model (ModelSEED)

Example KBase Workflow

1. Import FASTQ Data
   App: "Import FASTQ File as Reads"

2. Quality Control
   App: "Assess Read Quality with FastQC"

3. Trim Reads
   App: "Trim Reads with Trimmomatic"

4. Assemble
   App: "Assemble Reads with MEGAHIT"

5. Annotate
   App: "Annotate Assembly with PROKKA"

6. Bin
   App: "Bin Contigs with MetaBAT2"

7. Check Quality
   App: "Assess Genome Quality with CheckM"

8. Classify
   App: "Classify Genomes with GTDB-Tk"

9. Build Model
   App: "Build Metabolic Model with ModelSEED"

KBase Strengths

✅ Integrated platform (data + analysis + models)
✅ Metabolic modeling (unique feature!)
✅ Reproducible narratives
✅ Collaborative
✅ DOE compute resources

KBase Limitations

❌ Steeper learning curve
❌ Fewer tools than Galaxy
❌ US-focused (slower internationally)

🗄️ 3. IMG/M - Integrated Microbial Genomes & Microbiomes

IMG/M = JGI’s comprehensive database + analysis platform

Website: https://img.jgi.doe.gov

What is IMG/M?

Not just analysis - It’s a massive database!

• 300,000+ bacterial genomes
• 40,000+ metagenomes
• Pre-computed annotations
• Comparative tools

Using IMG/M

Upload Your Data

1. Create account (free)
2. Submit metagenome
3. Wait for processing (days-weeks)
4. Receive annotated results

Or Browse Public Data

Search by:

• Ecosystem
• Taxonomy
• Genes
• Functions
• Pathways

IMG/M Features

Automatic annotation includes:

• COG/KOG categories
• Pfam domains
• KEGG pathways
• Enzyme functions
• Signal peptides
• Transmembrane helices

Comparative tools:

• Gene phylogenetic profiler
• Abundance profiles
• Function comparisons
• BLAST searches

IMG/M Best For

✅ Comparing your data to public datasets
✅ Exploring gene families
✅ Metabolic pathway analysis
✅ Publication-quality annotations

❌ Not for quick turnaround
❌ Can’t customize pipeline

🦠 4. BV-BRC (formerly PATRIC) - Bacterial & Viral Database

BV-BRC = Bacterial and Viral Bioinformatics Resource Center

Website: https://www.bv-brc.org

BV-BRC Services

Genome Assembly

Services → Assembly
- Upload reads
- Choose assembler
- Submit job

Genome Annotation

Services → Annotation
- Upload genome
- Select taxonomy
- Choose annotation options
- Submit

Comprehensive Genome Analysis

Services → Comprehensive Genome Analysis
- Upload reads OR genome
- One-stop analysis:
  - Assembly
  - Annotation
  - Quality assessment
  - Specialized analyses

BV-BRC Analysis Tools

1. Proteome Comparison - Compare protein families
2. Phylogenetic Tree - Build trees from genomes
3. Protein Family Sorter - Find conserved proteins
4. Comparative Pathway Analysis - KEGG pathway comparison
5. Variation Analysis - SNP/indel identification
6. RNA-Seq Analysis - Transcriptomics
7. Tn-Seq Analysis - Transposon sequencing

BV-BRC Strengths

✅ Bacterial/viral focus
✅ AMR gene database (critical for pathogens)
✅ Drug target identification
✅ Extensive genome collection
✅ NIH-funded (reliable)

Best For

• Clinical/pathogen studies
• AMR research
• Vaccine development
• Outbreak analysis

🎯 Part III: Choosing the Right Tool

Decision Matrix

💡 Best Practices

For Workflow Wrappers

1. Start with default parameters
   • Don’t over-customize initially
   • Understand baseline before tweaking

2. Use version control

   git init my_project
   git add config.yaml
   git commit -m "Initial config"
   

3. Document your choices
   • Why this assembler?
   • Why these QC thresholds?
   • Keep a lab notebook 6
4. Test on subset
   • Don’t run 100 samples immediately
   • Test 1-3 samples first
5. Monitor resource usage

   htop  # Watch CPU/memory
   df -h # Check disk space
   

For Web Platforms

1. Start with tutorials
   • Galaxy has excellent training materials
   • KBase has guided narratives
2. Save workflows
   • Create reusable workflows
   • Share with collaborators
3. Check data limits
   • Know upload size limits
   • Understand storage quotas
4. Export results
   • Download key files locally
   • Don’t rely solely on web storage
5. Cite properly
   • Each platform has citation requirements
   • Acknowledge compute resources

🎓 Learning Resources

MetaWRAP

• Official docs
• Tutorial

nf-core/mag

• Pipeline docs
• Nextflow training

Anvi’o

• Official site
• Tutorials

Galaxy

• Training materials
• Metagenomics course

KBase

• Narratives
• Tutorials

✅ Success Checklist

• Chose appropriate workflow wrapper OR web platform
• Successfully ran test dataset
• Understood output structure
• Validated results against manual analysis (Days 1-7)
• Documented workflow/parameters
• Saved/exported results
• Ready to scale up to full dataset

📚 Key Takeaways

1. Workflow wrappers save time - Automate Days 1-7
2. Web platforms lower barriers - No Linux/HPC needed
3. Choose based on needs - No one-size-fits-all
4. Start simple - Use defaults before customizing
5. Reproducibility matters - Document everything

Repo for today’s code and other details

🔗 Day 8 →

Last updated: February 2026