Review and comparative analysis of machine learning-based phage virion protein identification methods

Area of Science:

• * Molecular Biology
• * Bioinformatics
• * Machine Learning

Background:

• * Phage virion protein (PVP) identification is essential for understanding phage-host interactions and designing biochemical entities.
• * Machine learning (ML) approaches have shown promise for PVP identification, necessitating systematic review and analysis.

Purpose of the Study:

• * To systematically review and analyze existing machine learning-based PVP identifiers.
• * To propose novel PVP identifiers based on a common framework.
• * To compare the performance of various PVP identifiers.

Main Methods:

• * Systematic review and analysis of existing PVP identification algorithms and tools.
• * Development of novel PVP identifiers utilizing a common framework.
• * Performance evaluation using training and independent datasets.

Main Results:

• * g-gap dipeptide composition (DPC) features effectively represent PVP characteristics.
• * Support vector machine (SVM) demonstrated superior performance in distinguishing PVPs from non-PVPs.
• * Comparison highlighted the strengths and weaknesses of different PVP identification methods.

Conclusions:

• * The study validates g-gap DPC features and SVM as effective tools for PVP identification.
• * The findings provide insights into optimizing ML approaches for PVP prediction.
• * This research contributes to advancing the understanding of phage biology and facilitating the design of related biotechnological applications.