gFACs

gFACs filters, analyzes, and converts predicted gene models and alignments to refine genome annotations and validate structural and functional gene features.

Key Features:

• Gene model filtering: Applies configurable filters to remove erroneous gene annotations based on structural criteria such as open reading frames, start sites, and splice sites.
• Model analysis: Computes and reports distributions and summary statistics of gene model attributes for assessment of the proposed gene space.
• File-format conversion: Converts predicted gene models and alignments among a range of bioinformatics file formats to enable integration of alignment, analysis, and gene prediction outputs.
• Support for long-read alignments: Handles text file formats produced by long read alignments and predicted gene structures to address errors arising from their integration.
• Functional evidence incorporation: Integrates functional attributes such as protein domains into gene model validation to retain models with verified structural and functional integrity.
• Flexible model definition: Provides a framework allowing user-defined structural and functional attributes for defining acceptable gene models.
• Filtering detail output: Produces detailed reports on filtering steps and distributions that can be visualized to further assess annotation quality.

Scientific Applications:

• Genome annotation refinement: Improves the quality and reliability of published genome annotations by removing structurally or functionally unsupported gene models.
• Gene model validation: Uses protein-domain evidence and structural criteria to validate predicted gene structures.
• Data integration: Facilitates integration of alignment and gene prediction results, including long-read alignment outputs, for downstream analyses.
• Preparation for genome browsers and downstream tools: Generates converted files and reports suitable for loading into genome browsers and other downstream workflows.

Methodology:

Implemented in Perl using BioPerl libraries, gFACs performs filtering, analysis, and file-format conversion of predicted gene models and alignments, incorporates protein-domain evidence, and outputs filtering statistics and distributions.