ACP_MS: prediction of anticancer peptides based on feature extraction

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Anticancer peptides (ACPs) show significant antitumor activity, making them promising cancer therapeutics.
• Accurate prediction of ACPs is crucial for advancing cancer research and drug development.

Purpose of the Study:

• To develop an efficient and accurate prediction model for anticancer peptides (ACPs).
• To enhance the identification capabilities for potential cancer treatment drugs.

Main Methods:

• Utilized the monoMonoKGap method for extracting digital features from anticancer peptide sequences.
• Employed the AdaBoost model for discriminating feature selection.
• Applied a stochastic gradient descent algorithm for ACP identification.
• Validated model performance using 7-fold cross-validation and independent test sets.

Main Results:

• The ACP_MS model achieved high accuracy on the main dataset (92.653% and 91.597%).
• Exceptional accuracy was observed on the alternate dataset (98.678% and 98.317%).
• ACP_MS demonstrated superior performance compared to existing advanced prediction models.

Conclusions:

• The developed ACP_MS model offers an improved method for identifying anticancer peptides.
• This tool facilitates further research into peptide-based cancer therapies.
• The model's data is publicly available for research use.