BUSCO Workflow¶

BUSCO (Benchmarking Universal Single-Copy Orthologs) provides quantitative assessment of genome assembly and annotation completeness based on evolutionarily-informed expectations.

Overview¶

BUSCO assesses genome quality by searching for conserved orthologous genes that are expected to be present in single copy in a given lineage. The presence, absence, or duplication of these genes provides insights into assembly and annotation quality.

How BUSCO Works¶

1. Lineage Selection: Choose an appropriate lineage dataset based on taxonomy
2. Gene Search: Search for conserved single-copy orthologs using HMMER
3. Classification: Genes are classified as Complete, Duplicated, Fragmented, or Missing
4. Scoring: Generate completeness scores and detailed reports

Analysis Modes¶

Protein Mode¶

Analyzes translated protein sequences from gene predictions.

When to use: - Assessing annotation quality - Evaluating gene prediction completeness - Comparing gene sets across species

Input: Protein FASTA from canonical transcripts

Advantages: - Direct assessment of annotation quality - Faster than genome mode - Better for comparing gene sets

Genome Mode¶

Analyzes the genome assembly directly using AUGUSTUS for gene prediction.

When to use: - Assessing assembly completeness - Checking for missing genes in annotation - Validating genome assembly quality

Input: Genome FASTA (DNA sequence)

Advantages: - Independent of existing annotation - Can detect genes missed by annotation pipeline - Better for assembly quality assessment

Both Modes¶

Run both protein and genome modes for comprehensive assessment.

When to use: - Complete quality control - Comparing annotation vs. assembly completeness - Production annotation validation

Provides: - Annotation completeness (protein mode) - Assembly completeness (genome mode) - Comparison of both metrics

Running BUSCO¶

Using Core Database¶

# Protein mode only
nextflow run main.nf \
  --csvFile genomes.csv \
  --run_busco_core \
  --busco_mode protein \
  --host mysql-server.example.com \
  --user_r ensro

# Genome mode only
nextflow run main.nf \
  --csvFile genomes.csv \
  --run_busco_core \
  --busco_mode genome \
  --host mysql-server.example.com \
  --user_r ensro

# Both modes (recommended for QC)
nextflow run main.nf \
  --csvFile genomes.csv \
  --run_busco_core \
  --busco_mode both \
  --host mysql-server.example.com \
  --user_r ensro

Using NCBI Assembly¶

# Automatically download and analyze genome from NCBI
nextflow run main.nf \
  --csvFile ncbi_genomes.csv \
  --run_busco_ncbi \
  --outdir results

Choosing a Lineage¶

Select the most specific lineage that applies to your organism. More specific lineages provide better sensitivity.

Decision Tree¶

Is your organism a...
├─ Eukaryote?
│  ├─ Animal (Metazoa)?
│  │  ├─ Vertebrate?
│  │  │  ├─ Mammal?
│  │  │  │  ├─ Primate? → primates_odb12
│  │  │  │  └─ Other → mammalia_odb12
│  │  │  ├─ Bird? → aves_odb12
│  │  │  ├─ Fish? → actinopterygii_odb12
│  │  │  └─ Other → vertebrata_odb12
│  │  └─ Invertebrate?
│  │     ├─ Insect?
│  │     │  ├─ Fly/Mosquito? → diptera_odb12
│  │     │  └─ Other → insecta_odb12
│  │     └─ Other → metazoa_odb12
│  ├─ Plant (Viridiplantae)?
│  │  ├─ Land plant? → embryophyta_odb12
│  │  └─ Other → viridiplantae_odb12
│  └─ Fungus?
│     ├─ Yeast? → saccharomycetes_odb12
│     └─ Other → fungi_odb12
└─ Bacteria? → bacteria_odb10

Common Lineages¶

Understanding Results¶

BUSCO Categories¶

BUSCO classifies each searched gene into one of four categories:

Score Interpretation¶

Completeness Score = (C + D) / Total BUSCOs × 100

Example Output¶

Short Summary (*_busco_short_summary.txt)¶

# BUSCO version is: 5.4.7 
# The lineage dataset is: primates_odb12 (Creation date: 2024-01-08, number of genomes: 44, number of BUSCOs: 13780)
# Summarized benchmarking in BUSCO notation for file /data/proteins.fa
# BUSCO was run in mode: protein
# Gene predictor used: None

***** Results: *****

C:98.5%[S:97.8%,D:0.7%],F:0.8%,M:0.7%,n:13780      
13570   Complete BUSCOs (C)            
13474   Complete and single-copy BUSCOs (S)    
96  Complete and duplicated BUSCOs (D)     
110 Fragmented BUSCOs (F)              
100 Missing BUSCOs (M)             
13780   Total BUSCO groups searched        

Interpretation: - ✅ Excellent completeness (98.5%) - ✅ Low duplication (0.7%) - ✅ Very few fragmented (0.8%) or missing (0.7%) - Conclusion: High-quality annotation

Full Table (*_busco_full_table.tsv)¶

# Busco id  Status  Sequence    Gene Start  Gene End    Strand  Score   Length
1000092at7742   Complete    ENSP00000262735 -   -   -   1256.4  416
1000093at7742   Complete    ENSP00000344818 -   -   -   2145.7  712
1000094at7742   Missing -   -   -   -   -   -
1000095at7742   Fragmented  ENSP00000123456 -   -   -   456.2   189

Output Files¶

Per-Sample Outputs¶

Directory Structure¶

results/
└── busco/
    ├── sample1_busco_short_summary.txt
    ├── sample1_genome_busco_short_summary.txt
    ├── sample1_busco_full_table.tsv
    ├── sample2_busco_short_summary.txt
    └── sample2_genome_busco_short_summary.txt

Interpreting Results by Mode¶

Protein vs. Genome Comparison¶

High Duplication (>5%)¶

Possible causes: - Polyploidy: Natural in some species (e.g., wheat, salmon) - Recent whole-genome duplication: Normal in some lineages - Misassembly: Collapsed repeats or chimeric contigs - Contamination: Mixed samples - Incorrect lineage: Using too broad a lineage

Action: Check genome characteristics and assembly quality metrics.

Troubleshooting¶

Low Completeness Scores¶

Problem: BUSCO completeness <80%

Possible causes: 1. Wrong lineage: Using inappropriate lineage dataset 2. Assembly gaps: Incomplete genome assembly 3. Annotation issues: Missing or incomplete gene predictions 4. Contamination: Non-target sequences in assembly 5. Unusual biology: Genuine gene loss (rare)

Solutions: - Try a broader lineage (e.g., vertebrata instead of primates) - Check assembly statistics (N50, gaps, contiguity) - Review annotation pipeline parameters - Screen for contamination

High Fragmentation¶

Problem: Many fragmented BUSCOs (>5%)

Possible causes: - Assembly fragmentation (low N50) - Incorrect gene models - Pseudogenes or processed pseudogenes

Solutions: - Improve assembly contiguity - Adjust gene prediction parameters - Check for frame shifts or premature stop codons

High Duplication¶

Problem: >10% duplicated BUSCOs (when not expected)

Possible causes: - Haplotigs not removed during assembly - Collapsed repeats - Chimeric sequences - Recent gene duplications (normal)

Solutions: - Run purge_dups or similar tools - Check assembly for haplotigs - Validate with Hi-C or genetic maps

Best Practices¶

Advanced Topics¶

Custom Lineages¶

For specialized projects, you can create custom lineage datasets:

1. Download from OrthoDB
2. Place in the lineage directory
3. Reference in your CSV or with --busco_dataset

Offline Mode¶

Pre-download lineage datasets:

# Download lineages
busco --download_path /data/busco_lineages --download vertebrata_odb12

# Use in pipeline
nextflow run main.nf \
  --csvFile genomes.csv \
  --run_busco_core \
  --download_path /data/busco_lineages

Metakey Integration¶

Load BUSCO results as database metakeys:

nextflow run main.nf \
  --csvFile genomes.csv \
  --run_busco_core \
  --apply_busco_metakeys \
  --host mysql-server.example.com \
  --user ensadmin \
  --password secret123

References¶

Primary Citations¶

• Manni et al. (2021). BUSCO Update: Novel and Streamlined Workflows along with Broader and Deeper Phylogenetic Coverage for Scoring of Eukaryotic, Prokaryotic, and Viral Genomes. Molecular Biology and Evolution, 38(10):4647–4654. DOI: 10.1093/molbev/msab199

• Manni et al. (2021). BUSCO: Assessing genomic data quality and beyond. Current Protocols, 1:e323. DOI: 10.1002/cpz1.323

Additional Resources¶

• BUSCO Website
• BUSCO User Guide
• OrthoDB Database
• BUSCO Publications

Next Steps¶

• OMArk Workflow - Complementary proteome assessment
• Output Documentation - Detailed output file specifications
• Troubleshooting - Solve common issues