PHPred

PHPred identifies bacteriophage proteins located within host cells from amino acid sequences to support functional characterization and antibacterial drug target discovery.

Key Features:

• Input data: Operates solely on amino acid sequence data.
• Computational methodology: Employs analysis of variance (ANOVA) in conjunction with incremental feature selection (IFS) to optimize feature sets.
• Classification performance: Achieves up to 84.2% overall accuracy in distinguishing bacteriophage proteins located within host cells from those not located there.
• Compartment-specific accuracy: Classifies proteins in specific cellular compartments—cytoplasm and membranes—with a maximum overall accuracy of 92.4%.
• Validation technique: Validated using jackknife cross-validation.

Scientific Applications:

• Functional understanding: Identifying bacteriophage proteins within host cells to aid elucidation of their roles and mechanisms during infection.
• Drug discovery: Supporting identification of potential targets for antibacterial drug development by highlighting phage proteins involved in bacterial infection processes.

Methodology:

ANOVA assesses variance among sequence-derived protein features, incremental feature selection (IFS) selects informative features, and performance is measured using jackknife cross-validation.