Mantis_GA

Mantis_GA performs consensus-driven protein function annotation by matching amino acid sequences with HMMER against multiple reference datasets and synthesizing results via text mining and a depth-first search to improve domain-specific annotation precision for genome and (meta-)omics studies.

Key Features:

• Consensus-Driven Annotation: Integrates annotations from diverse reference datasets using text mining to produce a consensus-driven annotation output.
• Flexibility and Customization: Allows full customization of reference datasets employed in the annotation process for tailored analyses.
• Speed and Efficiency: Annotates an average genome in approximately 25–40 minutes.
• High-Quality Annotations: Reports an average precision of 0.892 and coverage of 81.4%, employing a depth-first search to enhance domain-specific annotation relative to sequence-wide approaches.
• Reproducibility: Produces consistent outputs across different research contexts by integrating multiple reference sources into a consensus.

Scientific Applications:

• Genome-scale protein annotation: Functional assignment of proteins from individual organism genomes.
• (Meta-)omics and metagenomic studies: Functional annotation of proteins from community and environmental datasets at scale.
• Characterization of unannotated proteins: Inference of functional roles for previously unannotated proteins by consolidating evidence from well-characterized datasets.

Methodology:

Mantis_GA matches amino acid sequences to profile HMMs using HMMER, applies text mining to synthesize annotations from multiple reference datasets, and employs a depth-first search algorithm for domain-specific annotation.