TargetCLP: clathrin proteins prediction combining transformed and evolutionary scale modeling-based multi-view...

Area of Science:

• Molecular Biology
• Computational Biology
• Biochemistry

Background:

• Clathrin proteins are vital for cellular processes like endocytosis and neural function.
• Dysfunctional clathrin proteins are implicated in severe diseases including Alzheimer's, neurodegeneration, viral infections, and cancer.
• Accurate identification of clathrin proteins is essential for disease mechanism elucidation and drug target discovery.

Purpose of the Study:

• To develop a novel computational method, TargetCLP, for precise identification of clathrin proteins.
• To improve the accuracy and generalization capabilities of clathrin protein identification.
• To provide a tool for advancing research in diseases linked to clathrin dysfunction.

Main Methods:

• Leveraged four single-view feature representation methods: PSSM-CLBP, RECM-CLBP, qualitative characteristics, and ESM-based deep learning embeddings.
• Integrated features using differential evolution based on weights.
• Employed the BTG feature selection algorithm for optimal subset generation.
• Trained and evaluated models using various classifiers, notably the proposed SnBiLSTM.

Main Results:

• The TargetCLP method demonstrated significant improvements in prediction accuracy.
• The model exhibited enhanced generalization capabilities on independent datasets.
• The SnBiLSTM classifier achieved remarkable performance within the TargetCLP framework.

Conclusions:

• TargetCLP offers a significant advancement in the precise identification of clathrin proteins.
• The developed method provides a valuable tool for studying clathrin-related diseases.
• This work contributes to a deeper understanding of cellular mechanisms and potential therapeutic strategies.