Related Experiment Video
Updated: May 16, 2025
06:50
Author Spotlight: A Computational Approach to Decipher Amino Acid Preferences in Multispecific Protein-Protein Interactions
Published on: January 26, 2024
962
pACPs-DNN: Predicting anticancer peptides using novel peptide transformation into evolutionary and structure
Shahid1, Maqsood Hayat1, Ali Raza2
1Department of Computer Science, Abdul Wali Khan University Mardan, Mardan, KP 23200, Pakistan.
Computational Biology and Chemistry
|April 1, 2025
Summary
A new deep learning model, pACPs-DNN, accurately predicts anticancer peptides (ACPs) and non-ACPs. This computational tool enhances cancer drug discovery by identifying promising peptide therapeutics with high precision.
Area of Science:
- Computational biology
- Bioinformatics
- Drug discovery
Background:
- Cancer poses a significant global health burden with high mortality rates.
- Traditional cancer therapies are often resource-intensive and associated with severe side effects.
- Anticancer peptides (ACPs) offer a promising alternative due to their selectivity, safety, and ability to overcome drug resistance.
Purpose of the Study:
- To develop a novel deep learning model, pACPs-DNN, for accurate prediction of anticancer peptides (ACPs) and non-ACPs.
- To leverage advanced feature extraction and selection techniques for improved peptide classification.
- To provide a robust computational tool for accelerating the discovery of novel peptide-based cancer therapeutics.
Main Methods:
- Input peptides were transformed into image representations using residue-wise energy contact matrix (RECM), substitution Matrix Representation (SMR), and Position Specific Scoring Matrix (PSSM) embeddings.
- Local binary pattern (LBP)-based decomposition was applied to capture enhanced structural and local semantic features, generating novel feature sets (RECM_LBP, LBP_SMR, LBP_PSSM).
- A two-tier feature selection approach identified an optimal feature set for training an attention-based deep neural network.
Main Results:
- The pACPs-DNN model achieved a high training accuracy of 96.91% and an Area Under the Curve (AUC) of 0.98.
- Validation on independent datasets showed significant accuracy improvements of 5% and 3.5% over existing models on the Ind-I and Ind-II datasets, respectively.
- The model demonstrated strong generalization capability and robustness in predicting ACPs.
Conclusions:
- The pACPs-DNN model represents a significant advancement in the computational prediction of anticancer peptides.
- Its high accuracy and generalization performance highlight its potential as a valuable tool for drug discovery and cancer therapeutic development.
- This approach facilitates efficient identification of potential peptide candidates, aiding academic research and pharmaceutical development.
