Analysis, Modeling, and Target-Specific Predictions of Linear Peptides Inhibiting Virus Entry

Area of Science:

• Biochemistry
• Computational Biology
• Virology

Background:

• Antiviral peptides (AVPs) combat viruses via diverse mechanisms.
• Virus entry inhibitory peptides (VEIPs) specifically block enveloped viruses from infecting cells.
• Increasing VEIP data necessitates advanced prediction methods.

Purpose of the Study:

• To develop the first target-specific machine learning model for predicting novel VEIPs.
• To improve VEIP prediction by incorporating viral envelope protein attributes alongside peptide sequence characteristics.
• To address data scarcity for specific viral strains and enhance predictive power.

Main Methods:

• Developed a machine learning model integrating peptide sequence features and target virus envelope protein attributes.
• Utilized 10x10-fold cross-validation on a training dataset for model evaluation.
• Assessed model performance on an independent test set.

Main Results:

• The model achieved 87.33% accuracy and an MCC of 0.76 during cross-validation.
• On an independent test set, the model demonstrated 90.91% accuracy and an MCC of 0.81.
• A computational tool for automated VEIP prediction was created.

Conclusions:

• The developed model accurately predicts VEIPs, outperforming previous methods by integrating virus-specific information.
• This approach effectively overcomes data limitations for novel viral strains.
• The freely available tool facilitates accelerated discovery of potential antiviral therapeutics.