AGAT

AGAT processes and manipulates gene annotation files in GTF and GFF formats to manage, validate, and enrich genomic feature annotations for downstream analyses.

Key Features:

• Attribute Management: Relocates attributes within GTF/GFF records, allowing attributes to be moved upward or downward within feature entries.
• Functional Annotation Enhancement: Enriches annotations by importing UniProt fields (pe, sv, os) and BLAST metrics such as e-value into GTF/GFF records.
• Intron Analysis: Identifies introns shorter than expected via the script agat_sp_flag_short_introns_ebi.pl and flags records with the short_intron attribute.
• Warning Handling: Resolved warnings related to uninitialized values encountered during processing.
• Parser Fix for Consecutive Semicolons: Fixed parser behavior that previously inserted empty-string attributes when attributes were separated by consecutive semicolons.
• Isoform Testing Logic: Adjusted isoform testing to discard only failing isoforms unless all isoforms fail, in which case the parent gene is discarded.
• Empty ID List Fix: Corrected handling of empty ID lists that affected scripts such as agat_sp_filter_by_locus_distance.pl, agat_sp_fix_longest_ORF.pl, and agat_sp_filter_by_ORF_size.pl.

Scientific Applications:

• Annotation Curation: Manage and manipulate gene and transcript annotations in GTF/GFF formats for curation and quality control.
• Functional Annotation Augmentation: Augment annotations with UniProt-derived fields and BLAST e-values to support functional inference and annotation transfer.
• Intron Quality Control: Detect and flag unusually short introns for annotation correction and downstream filtering.
• Dataset Filtering: Filter and process large annotation datasets using locus-distance, ORF-size, and other criteria implemented in dedicated scripts.

Methodology:

Relocates attributes within records; parses and imports UniProt fields (pe, sv, os) and BLAST e-value into annotations; scans intron lengths and flags short introns with agat_sp_flag_short_introns_ebi.pl by adding the short_intron attribute; fixes include handling of uninitialized-value warnings, prevention of empty-string attributes from consecutive semicolons, adjusted isoform-testing logic, and corrected handling of empty ID lists in affected scripts.