StackCPPred: a stacking and pairwise energy content-based prediction of cell-penetrating peptides and their uptake...

Area of Science:

• Biotechnology
• Computational Biology
• Drug Delivery

Background:

• Cell-penetrating peptides (CPPs) are crucial for intracellular delivery of therapeutic molecules.
• Current experimental methods for CPP identification are costly and time-consuming.
• Computational prediction of CPPs can accelerate research and development.

Purpose of the Study:

• To develop an advanced computational method for predicting CPPs.
• To improve feature representation for enhanced CPP prediction accuracy.
• To provide a tool for efficient identification of CPPs for therapeutic applications.

Main Methods:

• Proposed StackCPPred, a stacking-based machine learning approach.
• Utilized three novel feature representation methods: RECM-composition, PseRECM, and RECM-DWT.
• Employed jackknife validation on CPP924 and CPPsite3 datasets.

Main Results:

• StackCPPred achieved 94.5% accuracy on the CPP924 dataset and 78.3% on the CPPsite3 dataset.
• Demonstrated significant performance improvement over existing state-of-the-art CPP predictors.
• Indicated the potential of StackCPPred in predicting CPPs and their uptake efficiency.

Conclusions:

• StackCPPred offers a powerful and efficient computational tool for CPP prediction.
• The method's improved feature representation addresses limitations of previous approaches.
• Facilitates hypothesis-driven experimental design and accelerates CPPs' clinical therapy applications.