bio-genome-annotation-functional-annotation

name	bio-genome-annotation-functional-annotation
description	Assign GO terms, KEGG orthologs, Pfam domains, and EC numbers to predicted proteins using eggNOG-mapper and InterProScan. Produces functional summaries for downstream pathway and enrichment analysis. Use when adding functional annotation to predicted genes or characterizing protein functions in a new genome.
tool_type	cli
primary_tool	eggNOG-mapper

Functional Annotation

Assign functional annotations (GO terms, KEGG orthologs, Pfam domains, EC numbers) to predicted protein sequences using eggNOG-mapper and InterProScan.

eggNOG-mapper

Database Setup

# Download eggNOG v5.0 database (~44 GB)
# Required for local searches; use --data_dir to specify location
download_eggnog_data.py --data_dir /path/to/eggnog_db -y

# Download DIAMOND database only (~9 GB, faster setup)
download_eggnog_data.py --data_dir /path/to/eggnog_db -y -D

# Download taxon-specific databases (optional, smaller)
download_eggnog_data.py --data_dir /path/to/eggnog_db -y -t 2 # Bacteria
download_eggnog_data.py --data_dir /path/to/eggnog_db -y -t 2759 # Eukaryota

Basic Usage

emapper.py \
    -i predicted_proteins.faa \
    --output functional_annot \
    --output_dir eggnog_out \
    --data_dir /path/to/eggnog_db \
    --cpu 16 \
    -m diamond

Key Options

Option	Description
-i	Input protein FASTA
--output	Output file prefix
--data_dir	Path to eggNOG database
-m	Search mode: diamond (fast), mmseqs (sensitive), hmmer
--cpu	CPU threads
--tax_scope	Taxonomic scope (auto, Bacteria, Eukaryota, etc.)
--go_evidence	GO evidence filter (experimental, non-electronic, all)
--target_orthologs	Ortholog type (one2one, all)
--seed_ortholog_evalue	E-value cutoff (default: 0.001)
--seed_ortholog_score	Min bit score (default: 60)
--override	Overwrite existing output

With Taxonomic Scope

# Restrict to bacterial orthologs for a prokaryotic genome
emapper.py \
    -i proteins.faa \
    --output annot \
    --output_dir eggnog_out \
    --data_dir /path/to/eggnog_db \
    --cpu 16 \
    -m diamond \
    --tax_scope Bacteria \
    --go_evidence non-electronic

Output Files

eggnog_out/
├── annot.emapper.annotations    # Main annotation table
├── annot.emapper.hits           # DIAMOND/mmseqs hits
├── annot.emapper.seed_orthologs # Best orthologs
└── annot.emapper.pfam           # Pfam domain annotations

Key Output Columns

Column	Content
seed_ortholog	Best matching ortholog
evalue	E-value of best hit
GOs	GO term annotations
EC	Enzyme Commission numbers
KEGG_ko	KEGG ortholog IDs
KEGG_Pathway	KEGG pathway mappings
COG_category	COG functional category
PFAMs	Pfam domain annotations
Description	Functional description

InterProScan

InterProScan searches multiple protein signature databases simultaneously.

Basic Usage

interproscan.sh \
    -i predicted_proteins.faa \
    -o interpro_results.tsv \
    -f tsv,gff3 \
    -cpu 16 \
    -goterms \
    -pa

Key Options

Option	Description
-i	Input protein FASTA
-o	Output file
-f	Output formats: tsv, gff3, xml, json
-cpu	CPU threads
-goterms	Include GO term mappings
-pa	Include pathway annotations
-appl	Specific applications to run (comma-separated)
-dp	Disable precalculated match lookup

Select Specific Databases

# Run only Pfam, TIGRFAM, and CDD
interproscan.sh \
    -i proteins.faa \
    -o interpro_results.tsv \
    -f tsv,gff3 \
    -cpu 16 \
    -goterms -pa \
    -appl Pfam,TIGRFAM,CDD

Available Applications

Application	Description
Pfam	Protein families
TIGRFAM	Functionally equivalent protein families
SUPERFAMILY	Structural domain assignments
CDD	Conserved Domain Database
PANTHER	Protein classification
Gene3D	Structural domain predictions
Coils	Coiled-coil predictions
MobiDBLite	Disordered regions
SignalP	Signal peptides
TMHMM	Transmembrane helices

Merging eggNOG and InterProScan Results

import pandas as pd

def parse_eggnog(annotations_file):
    '''Parse eggNOG-mapper annotations output.'''
    df = pd.read_csv(annotations_file, sep='\t', comment='#',
                     header=None, skiprows=5)
    col_names = [
        'query', 'seed_ortholog', 'evalue', 'score', 'eggNOG_OGs',
        'max_annot_lvl', 'COG_category', 'Description', 'Preferred_name',
        'GOs', 'EC', 'KEGG_ko', 'KEGG_Pathway', 'KEGG_Module',
        'KEGG_Reaction', 'KEGG_rclass', 'BRITE', 'KEGG_TC', 'CAZy',
        'BiGG_Reaction', 'PFAMs'
    ]
    df.columns = col_names[:len(df.columns)]
    return df

def parse_interproscan_tsv(tsv_file):
    '''Parse InterProScan TSV output.'''
    col_names = [
        'protein_id', 'md5', 'length', 'analysis', 'signature_acc',
        'signature_desc', 'start', 'stop', 'score', 'status', 'date',
        'interpro_acc', , , 
    ]
    df = pd.read_csv(tsv_file, sep=, header=, names=col_names)
     df

 ():
    
    eggnog_df = parse_eggnog(eggnog_file)
    interpro_df = parse_interproscan_tsv(interpro_file)

    interpro_summary = interpro_df.groupby().agg({
        :  x: .join(x.dropna().unique()),
        :  x: .join(x.dropna().unique()),
        :  x: .join(x.dropna().unique()),
    }).reset_index()
    interpro_summary.columns = [, , , ]

    merged = eggnog_df.merge(interpro_summary, on=, how=)

    merged[] = merged.apply(
         row: combine_go_terms(row.get(, ), row.get(, )), axis=
    )
     merged

 ():
    
    terms = ()
     go_str  [eggnog_go, interpro_go]:
         pd.notna(go_str)  go_str != :
            terms.update(t.strip()  t  (go_str).replace(, ).split()  t.strip().startswith())
     .join((terms))  terms  

Annotation Statistics

def annotation_summary(merged_df):
    '''Summarize functional annotation coverage.'''
    total = len(merged_df)
    has_go = (merged_df['all_go'] != '-').sum()
    has_kegg = merged_df['KEGG_ko'].notna().sum() if 'KEGG_ko' in merged_df else 0
    has_pfam = merged_df['PFAMs'].notna().sum() if 'PFAMs' in merged_df else 0
    has_ec = merged_df['EC'].notna().sum() if 'EC' in merged_df else 0
    has_desc = (merged_df['Description'] != '-').sum() if 'Description' in merged_df else 0

    print(f'Total proteins: {total}')
    print(f'With GO terms: {has_go} ({has_go/total:.1%})')
    print(f'With KEGG orthologs: {has_kegg} ({has_kegg/total:.1%})')
    print(f'With Pfam domains:  ()')
    ()
    ()

    
    has_any = ((merged_df[] != ) | merged_df[].notna() | merged_df[].notna()).()
    ()

Troubleshooting

Low Annotation Rate

• Check protein sequence quality (no fragmented ORFs)
• Try broader taxonomic scope (--tax_scope auto)
• Run both eggNOG-mapper and InterProScan and merge results

eggNOG Database Errors

• Verify database version matches emapper version
• Re-download with download_eggnog_data.py --data_dir /path -y

InterProScan Memory Issues

• Reduce batch size with -b option
• Split input FASTA into smaller chunks

Related Skills

• prokaryotic-annotation - Bakta includes basic functional annotation
• eukaryotic-gene-prediction - Produces protein sequences for functional annotation
• pathway-analysis/go-enrichment - Enrichment analysis using GO annotations
• pathway-analysis/kegg-pathways - Pathway mapping with KEGG orthologs