TiCo

TiCo refines translation initiation site (TIS) predictions in prokaryotic genomes to improve localization of gene start sites for genome annotation and comparative analyses.

Key Features:

• Unsupervised clustering algorithm: Employs an unsupervised clustering algorithm that operates without prior assumptions about prokaryotic TIS characteristics, enabling adaptation to diverse genomic contexts.
• Positionally smoothed probability matrices: Uses positionally smoothed probability matrices to score potential TIS locations, improving robustness in genomes with high G+C content.
• GLIMMER post-processing: Functions as a post-processor of GLIMMER outputs to refine initial gene predictions and localize TIS positions more precisely.
• Versatility across genomes: Demonstrated competitive performance on experimentally verified test data from multiple bacterial species, including Escherichia coli, Bacillus subtilis, Pseudomonas aeruginosa, Burkholderia pseudomallei, and Ralstonia solanacearum.
• Robustness to annotation variation: Maintains reliable performance despite variations in initial TIS annotations without requiring specific assumptions about gene starts.

Scientific Applications:

• Genome annotation: Improves accuracy of gene start site annotations in prokaryotic genome annotation projects.
• Comparative genomics: Enhances cross-species comparisons by providing more consistent TIS localization across bacterial genomes.
• Evolutionary and regulatory studies: Supports analyses of regulatory mechanisms and evolutionary patterns that depend on accurate TIS identification.

Methodology:

TiCo applies an unsupervised clustering algorithm combined with positionally smoothed probability matrices to score candidate TIS positions as a post-processing step for GLIMMER-predicted genes.