iTerm-PseKNC

iTerm-PseKNC predicts transcriptional terminators in bacterial genomes to support identification of operon structures and improve genome annotation.

Key Features:

• Support Vector Machine (SVM): Used as the classification framework to identify transcriptional terminators.
• Pseudo k-tuple nucleotide composition (PseKNC): Extracts features that capture sequence order and nucleotide physicochemical properties.
• Binomial distribution feature selection: Applies a binomial distribution approach to optimize the feature subset for the predictive model.
• Rho-independent terminator prediction: Targets prediction of experimentally confirmed Rho-independent transcriptional terminators.
• Cross-validation evaluation: Assessed using 5-fold cross-validation, reporting 95% accuracy.
• Independent-dataset validation: Validated on experimentally confirmed Rho-independent terminators from Escherichia coli and Bacillus subtilis, identifying all E. coli terminators and 87.5% of B. subtilis terminators.

Scientific Applications:

• Transcription regulation studies: Enables analysis of transcription termination events and regulatory mechanisms.
• Operon structure determination: Aids identification of operon boundaries by locating terminators.
• Genome annotation: Enhances genome annotation by adding predicted transcriptional terminator locations.

Methodology:

Modeling used a support vector machine with features derived from PseKNC, feature subset optimization via a binomial distribution approach, evaluation by 5-fold cross-validation, and validation on independent datasets of experimentally confirmed Rho-independent terminators from Escherichia coli and Bacillus subtilis.