TIGR Software Tools

TIGR Software Tools provides a suite for ab initio gene prediction in eukaryotic genomes using GHMM, HMM, decision-tree, and Interpolated Markov Model algorithms to identify protein-coding genes.

Key Features:

• Exonomy: Implements a 23-state Generalized Hidden Markov Model (GHMM) for eukaryotic gene prediction.
• Unveil: Implements a 283-state standard Hidden Markov Model (HMM) for detailed state modeling in gene prediction.
• GlimmerM: Combines decision trees with Interpolated Markov Models (IMMs) to predict genes.
• Retrainability: Models can be retrained for new organisms to adapt to diverse genomic datasets.
• Ensemble usage: Individual programs can be combined in ensemble approaches to improve prediction performance under specific conditions.

Scientific Applications:

• Eukaryotic genomics: Applied to ab initio identification of protein-coding genes in eukaryotic genomes.
• Model organism validation: Methods were validated using Arabidopsis thaliana as a test organism.
• Comparative prediction: Different algorithms can outperform others under specific conditions, supporting comparative and ensemble analyses.
• Cross-species adaptation: Retrainability enables application across diverse eukaryotic species.

Methodology:

Exonomy: 23-state GHMM; Unveil: 283-state HMM; GlimmerM: decision trees with Interpolated Markov Models (IMMs); model retraining for new organisms.