Therapeutic peptides identification via kernel risk sensitive loss-based k-nearest neighbor model and multi-Laplacian...

Area of Science:

• Biotechnology
• Computational Biology
• Drug Discovery

Background:

• Therapeutic peptides, synthesized from amino acids, offer precise drug delivery and immune system activation for disease treatment.
• Traditional biochemical assays for screening therapeutic peptides are costly, time-consuming, and face experimental and ethical limitations.
• Machine learning and computational methods present an efficient, automated alternative for predicting therapeutic peptides.

Purpose of the Study:

• To develop and evaluate a novel machine learning model for predicting therapeutic peptides.
• To address the limitations of conventional screening methods through computational approaches.

Main Methods:

• Proposed a k-nearest neighbor (k-NN) model incorporating multi-Laplacian regularization.
• Introduced a kernel risk-sensitive loss function derived from the K-local hyperplane distance nearest neighbor model.
• Utilized computational methods for automated prediction of therapeutic peptide sequences.

Main Results:

• The proposed k-NN model demonstrated satisfactory performance in predicting therapeutic peptides.
• The approach effectively identified potential therapeutic peptide sequences.
• The computational method proved efficient and accurate compared to traditional assays.

Conclusions:

• The developed machine learning model offers a viable and effective computational strategy for therapeutic peptide discovery.
• This method enhances the efficiency and accuracy of identifying potential therapeutic peptides, overcoming experimental and ethical challenges.
• The study highlights the potential of advanced machine learning techniques in accelerating drug discovery processes.