Whole-Genome Annotation with BRAKER.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate gene prediction is crucial for understanding eukaryotic genome function.
• Existing gene prediction tools often require manual parameterization or lack flexibility in integrating diverse evidence types.

Purpose of the Study:

• To present BRAKER, a highly accurate and fully automated pipeline for eukaryotic gene prediction.
• To describe the methodology and application of BRAKER with various combinations of extrinsic evidence.

Main Methods:

• BRAKER combines GeneMark-ES/ET for automated parameter learning and AUGUSTUS for accurate gene finding.
• It integrates extrinsic evidence, including RNA-Seq read footprints, cross-species protein alignments, and heterogeneous data.
• The pipeline refines models using available evidence during both training and prediction steps.

Main Results:

• BRAKER achieves highly accurate and fully automated gene prediction in novel eukaryotic genomes.
• The pipeline demonstrates flexibility in utilizing diverse extrinsic evidence, enhancing prediction robustness.
• Previous versions like BRAKER1 established foundational integration of RNA-Seq data.

Conclusions:

• BRAKER provides a robust and automated solution for gene prediction in eukaryotic genomes.
• Its ability to integrate heterogeneous evidence makes it adaptable to various genomic research scenarios.
• The pipeline facilitates efficient and accurate analysis of novel eukaryotic genomes.