Glimmer

Glimmer identifies genes in microbial DNA (bacteria, archaea, and viruses) using interpolated Markov models (IMMs) to distinguish coding regions from noncoding DNA.

Key Features:

• Interpolated Markov Models (IMMs): Employs IMMs that capture dependencies between nearby nucleotides with variable context lengths to model coding versus noncoding sequence composition.
• High Sensitivity and Accuracy: Reports approximately 99% sensitivity in detecting genes across most species and recovers roughly 97–98% of genes compared with published annotations, missing under 1% of known genes with significant homology.
• Improved Start Codon Identification: Incorporates enhanced methods for identifying start codons, improving correct identification of gene start sites relative to curated genes.
• Distinguishing Host from Endosymbiont DNA: Includes a module to differentiate host from endosymbiont DNA to reduce false-positive bacterial gene predictions in eukaryotic sequencing projects.
• Comprehensive Evaluation and Technical Improvements: Has undergone extensive evaluations and technical improvements that increased precision in gene identification.

Scientific Applications:

• Microbial genome annotation: Generates gene predictions for bacterial, archaeal, and viral genomes to support genome annotation efforts.
• Functional annotation and evolutionary studies: Provides coding region and start site predictions used in functional annotation and evolutionary biology analyses.
• Complex sequencing projects: Enables separation of host and endosymbiont sequences in eukaryotic genome projects to mitigate bacterial contamination in downstream analyses.

Methodology:

Uses interpolated Markov models (IMMs) to model nucleotide dependencies and distinguish coding from noncoding regions, applies enhanced algorithms for start codon identification, and employs a module to differentiate host versus endosymbiont DNA, supported by extensive evaluations and technical improvements.