Artificial neural network study on organ-targeting peptides.

Area of Science:

• Computational biology
• Peptide science
• Drug discovery

Background:

• Identifying organ-targeting peptides is crucial for developing targeted therapeutics.
• Current methods for peptide discovery can be time-consuming and resource-intensive.

Purpose of the Study:

• To develop and validate a novel computational approach for predicting organ-targeting peptides based on their amino acid sequences.
• To assess the predictive power of machine learning models using VHSE descriptors and simple neural network architectures.

Main Methods:

• Utilized positive control datasets of organ-targeting peptides identified via peroral phage display for four organs.
• Generated negative control datasets from random sequences.
• Employed VHSE descriptors and simple neural network architectures for model training and validation.
• Evaluated model performance using statistical indicators such as sensitivity, specificity, and the area under the receiver operating characteristic (ROC) curve.

Main Results:

• VHSE descriptor produced statistically significant training models.
• Models with simple neural network architectures demonstrated slightly greater predictive power than complex ones.
• The developed models effectively discriminated between organ-targeting and random peptide sequences.
• Statistical indicators confirmed the models' capacity for accurate predictions.

Conclusions:

• The novel computational approach accurately predicts organ-targeting peptides from sequence information.
• Simple neural network architectures offer efficient and effective prediction of peptide organotropism.
• This method holds promise for the selection of peptides in the development of peptide drugs and peptidomimetics.