Genome Annotation
Annotation Track • Chapter 5
Estimated Study Time: 20 minutes
Learning Objectives
- Distinguish structural annotation (gene coordinate boundaries) from functional annotation (biological roles).
- Execute de novo repeat identification and soft-masking using RepeatModeler and RepeatMasker.
- Train ab initio gene predictors (Augustus, GeneMark) and combine extrinsic RNA-seq/protein evidence via BRAKER3.
- Assign functional gene ontologies, PFAM domains, and EC numbers using InterProScan and eggNOG-mapper.
1. Structural vs Functional Genome Annotation
Once a draft or chromosome-scale genome assembly is constructed, it consists merely of millions of letters of contiguous sequence (FASTA). Genome annotation is the multi-stage computational process of attaching biological meaning to these coordinates:
- Structural Annotation: Delineating the precise coordinate locations of genomic features: protein-coding genes (exons, introns, 5' UTR, 3' UTR), non-coding RNAs (tRNA, rRNA, snRNA, lncRNA), and transposable elements. Output is stored in 9-column GFF3 (Generic Feature Format) or GTF files.
- Functional Annotation: Assigning biological identities, protein domain architectures, catalytic functions, and cellular locations to the predicted protein sequences by searching curated databases (UniProt, Pfam, InterPro, Gene Ontology, KEGG).
2. BRAKER3 & eggNOG-mapper Annotation Pipeline
# Step 1: De novo repeat modeling and soft-masking
BuildDatabase -name my_genome_db -engine ncbi draft_scaffolds.fasta
RepeatModeler -database my_genome_db -pa 16 -LTRStruct
RepeatMasker -pa 16 -lib my_genome_db-families.fa -xsmall draft_scaffolds.fasta
# Step 2: Automated gene prediction with BRAKER3 combining RNA-seq and protein evidence
braker.pl --genome=draft_scaffolds.fasta.masked \
--bam=spliced_rnaseq.bam \
--prot_seq=vertebrate_orthologs.fa \
--threads=16 --gff3 --species=my_organism
# Step 3: Functional annotation with eggNOG-mapper
emapper.py -i braker_proteins.faa --output eggnog_results \
--cpu 16 -m diamond --sensmode more-sensitive \
--data_dir /path/to/eggnog_db/