DLBWE-Cys: a deep-learning-based tool for identifying cysteine S-carboxyethylation sites using binary-weight...

Area of Science:

• Biochemistry
• Computational Biology
• Immunology

Background:

• Cysteine S-carboxyethylation is a novel post-translational modification (PTM) implicated in autoimmune disease pathogenesis, such as ankylosing spondylitis.
• Accurate identification of S-carboxyethylation sites is crucial for understanding its functional roles.
• Existing computational tools lack the accuracy needed for predicting these modification sites, hindering research progress.

Purpose of the Study:

• To develop a novel deep learning model for accurate prediction of cysteine S-carboxyethylation modification sites.
• To address the current limitations in computational tools for identifying these PTMs.
• To provide a valuable resource for researchers studying autoimmune diseases.

Main Methods:

• Developed DLBWE-Cys, a deep learning model integrating CNN, BiLSTM, Bahdanau attention, and FNN.
• Utilized Binary-Weight encoding specifically designed for cysteine S-carboxyethylation site prediction.
• Evaluated model performance using 5-fold cross-validation and independent testing, including t-SNE visualization.

Main Results:

• DLBWE-Cys demonstrated superior performance compared to existing machine learning and deep learning models.
• Feature comparison experiments confirmed the effectiveness of Binary-Weight encoding over other methods.
• t-SNE visualization validated the model's robust classification capabilities.

Conclusions:

• The developed DLBWE-Cys model offers a highly accurate method for predicting cysteine S-carboxyethylation sites in protein sequences.
• This tool is expected to significantly advance research into the functional mechanisms of S-carboxyethylation in diseases.
• The model and data are publicly available, facilitating further research and development.