GPRED-GC

GPRED-GC predicts protein-coding genes in genomes with highly variable GC content using a hidden Markov model to detect genes exhibiting complex GC patterns, including negative 5'-3' GC gradients common in grass genomes.

Key Features:

• Ab initio Gene Prediction: Performs gene prediction without reliance on homologous sequences to enable annotation of new genomes.
• Hidden Markov Model (HMM): Uses an HMM framework to model intragenic GC content variation within predicted gene models.
• Optimization for Variable GC Content: Specifically optimized to detect genes with highly variable GC content, including negative 5'-3' GC gradients.
• Complementary to Existing Tools: Designed to complement gene predictors such as Augustus to improve sensitivity and accuracy in gene discovery for GC-variable genomes.

Scientific Applications:

• Genome annotation: Ab initio annotation of genomes, particularly those lacking homologous reference sequences.
• Grass genome analysis: Identification of genes exhibiting negative 5'-3' GC gradients in grasses.
• Model plant benchmarking: Application to Arabidopsis thaliana and Oryza sativa datasets to detect genes with highly variable GC content.

Methodology:

Implements an HMM-based ab initio gene prediction approach that models intragenic GC content variation to identify genes with complex GC profiles.