Prediction of anticancer peptides based on an ensemble model of deep learning and machine learning using ordinal...

Area of Science:

• Biotechnology
• Computational Biology
• Bioinformatics

Background:

• Anticancer peptides (ACPs) show promise as safer, more selective cancer treatment alternatives.
• Current ACP identification methods are costly, time-consuming, and lab-intensive.
• Developing efficient computational prediction tools for ACPs is crucial.

Purpose of the Study:

• To develop and validate a novel computational method for predicting anticancer peptides (ACPs) using sequence information.
• To improve the efficiency and reduce the cost associated with ACP identification.
• To provide a valuable tool for cancer research and drug discovery.

Main Methods:

• A hybrid computational model integrating deep learning (bidirectional long short-term memory and convolutional neural network) and machine learning (Light Gradient Boosting Machine) was developed.
• Peptide sequences were processed using ordinal encoding with positional information and handcrafted features.
• A model ensemble approach, combining results from deep learning and machine learning modules, was employed for final classification.

Main Results:

• The developed model achieved high performance metrics: 0.7895 accuracy, 0.8153 sensitivity, and 0.7676 specificity.
• The model demonstrated at least a 2% improvement across all metrics compared to existing state-of-the-art methods.
• The study utilized a benchmark dataset, confirming the model's effectiveness and potential for further enhancement.

Conclusions:

• The proposed computational method offers a more effective and efficient approach for predicting anticancer peptides.
• This novel technique can accelerate the discovery of potential ACPs for cancer prevention and treatment.
• The source code and research are publicly available to facilitate community-driven advancements in ACP research.