The NCBI Prokaryotic Genome Annotation Pipeline (PGAP) is an advanced tool for automatic annotation of prokaryotic genomes, combining alignment-based methods with ab initio gene prediction. The pipeline uses a pan-genome approach, leveraging homologous protein clusters to generate initial annotations and refine them using statistical predictions. Key features include: 1. **Pan-genome Approach**: Utilizes homologous protein clusters to define core genes and proteins, which serve as a map for annotation. 2. **Multiple Evidence Types**: Integrates alignment-based hints and ab initio predictions to improve accuracy. 3. **Two-Pass Annotation**: Aims to detect frameshifted genes and pseudogenes, enhancing the robustness of annotations. 4. **GeneMarkS+**: A self-training gene finder that integrates external evidence to predict protein-coding regions. 5. **High-Performance Execution**: Uses a modular framework (GPipe) for distributed computing and robust tracking of tasks. 6. **Quality Control**: Incorporates validation procedures to ensure biologically valid and consistently formatted data. The PGAP pipeline has been shown to match GenBank annotations in over 98% of cases and has been integrated into the GenBank submission system. It has also been used to re-annotate RefSeq genomes, improving consistency and accuracy. The pipeline is designed to handle both well-studied and novel taxonomic lineages, making it a flexible and extensible tool for prokaryotic genome annotation.The NCBI Prokaryotic Genome Annotation Pipeline (PGAP) is an advanced tool for automatic annotation of prokaryotic genomes, combining alignment-based methods with ab initio gene prediction. The pipeline uses a pan-genome approach, leveraging homologous protein clusters to generate initial annotations and refine them using statistical predictions. Key features include: 1. **Pan-genome Approach**: Utilizes homologous protein clusters to define core genes and proteins, which serve as a map for annotation. 2. **Multiple Evidence Types**: Integrates alignment-based hints and ab initio predictions to improve accuracy. 3. **Two-Pass Annotation**: Aims to detect frameshifted genes and pseudogenes, enhancing the robustness of annotations. 4. **GeneMarkS+**: A self-training gene finder that integrates external evidence to predict protein-coding regions. 5. **High-Performance Execution**: Uses a modular framework (GPipe) for distributed computing and robust tracking of tasks. 6. **Quality Control**: Incorporates validation procedures to ensure biologically valid and consistently formatted data. The PGAP pipeline has been shown to match GenBank annotations in over 98% of cases and has been integrated into the GenBank submission system. It has also been used to re-annotate RefSeq genomes, improving consistency and accuracy. The pipeline is designed to handle both well-studied and novel taxonomic lineages, making it a flexible and extensible tool for prokaryotic genome annotation.