GeMoMa

GeMoMa predicts protein-coding gene models in target genomes by homology to annotated reference genomes, using amino acid sequence and intron position conservation and optional RNA-seq evidence.

Key Features:

• Homology-based projection: Uses annotated protein-coding genes from a reference genome to predict annotations in a target genome.
• Amino acid and intron conservation: Leverages amino acid sequence conservation and intron position conservation to improve gene model accuracy.
• RNA-seq integration: Incorporates RNA-seq data for splice site prediction and to provide experimental evidence for gene models.
• Multiple reference organisms: Supports use of multiple reference genomes to broaden comparative evidence for predictions.
• Comparative performance: Benchmarked against BRAKER1, MAKER2, CodingQuarry, and purely RNA-sequencing-based pipelines, with reported superior performance in evaluations.
• Homolog detection: Identifies homologs of specific genes or gene families and refines protein-coding gene annotations.

Scientific Applications:

• Genome annotation: Annotates newly sequenced genomes and refines existing protein-coding gene annotations.
• Homolog identification: Detects homologs of specific genes or gene families across species.
• Empirical use cases: Applied to annotate four nematode species and the barley genome.
• Cross-taxa evaluation: Used in benchmarking and evaluations across plants, animals, and fungi.

Methodology:

Performs homology-based projection of annotated protein-coding genes from single or multiple reference genomes onto a target genome using amino acid sequence conservation and intron position conservation, with optional incorporation of RNA-seq evidence for splice site prediction.