TargetCPP: accurate prediction of cell-penetrating peptides from optimized multi-scale features using gradient boost...

Area of Science:

• Bioinformatics
• Molecular Biology
• Drug Discovery

Background:

• Cell-penetrating peptides (CPPs) are vital for delivering therapeutic agents into cells.
• Accurate CPP identification is essential for advancing medical applications and disease treatment.
• Traditional experimental methods for CPP identification are time-consuming and have limitations.

Purpose of the Study:

• To develop an accurate computational method for identifying cell-penetrating peptides (CPPs).
• To overcome the limitations of experimental CPP identification methods.
• To provide a tool for large-scale prediction of CPPs from uncharacterized sequences.

Main Methods:

• Developed TargetCPP, a novel computational method for CPP discrimination.
• Utilized four distinct peptide sequence encoding methods.
• Applied feature selection (minimum redundancy and maximum relevancy) and gradient boost decision tree classification.

Main Results:

• TargetCPP achieved high prediction accuracy: 93.54% (jackknife) and 88.28% (independent test).
• The gradient boost decision tree classifier demonstrated excellent performance.
• The proposed method shows superiority over existing state-of-the-art techniques.

Conclusions:

• TargetCPP offers a potent and accurate bioinformatics approach for CPP identification.
• This tool can support large-scale CPP prediction and guide clinical therapy.
• The findings provide a strong foundation for future medical applications of CPPs.