Scholarlink

Protein ubiquitination is a critical post-translational modification involved in numerous biological processes, playing a vital role in regulating both physiological and pathological mechanisms. Despite existing tools for predicting ubiquitination sites, variations in ubiquitination patterns among species remain inadequately understood. This study presents LBiPlantUbi, a novel hybrid deep learning model that incorporates Long Short-Term Memory (LSTM) and Bidirectional LSTM (Bi-LSTM) networks combined with natural language processing (NLP) techniques for predicting ubiquitination sites in plants. Our model uniquely learns features directly from raw protein sequences using word embedding and integrates feature extraction and prediction within a single architecture. LBiPlantUbi achieved superior performance with an accuracy of 84%, sensitivity of 88.9%, F1-score of 0.844, MCC of 0.675, and AUC of 0.902 on independent testing sets, outperforming existing methods. This advanced model opens new avenues for studying ubiquitination in plants and other species. The implementation is publicly accessible on GitHub at https://github.com/nuinvtnu/LBiPlantUbi .

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Hybrid Deep Learning and Natural Language Processing Model for Plant Ubiquitination Sites Prediction

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Thi-Xuan Tran,
  • Thi-Tuyen Nguyen,
  • Nguyen Quoc Khanh Le,
  • Van-Nui Nguyen

摘要

Protein ubiquitination is a critical post-translational modification involved in numerous biological processes, playing a vital role in regulating both physiological and pathological mechanisms. Despite existing tools for predicting ubiquitination sites, variations in ubiquitination patterns among species remain inadequately understood. This study presents LBiPlantUbi, a novel hybrid deep learning model that incorporates Long Short-Term Memory (LSTM) and Bidirectional LSTM (Bi-LSTM) networks combined with natural language processing (NLP) techniques for predicting ubiquitination sites in plants. Our model uniquely learns features directly from raw protein sequences using word embedding and integrates feature extraction and prediction within a single architecture. LBiPlantUbi achieved superior performance with an accuracy of 84%, sensitivity of 88.9%, F1-score of 0.844, MCC of 0.675, and AUC of 0.902 on independent testing sets, outperforming existing methods. This advanced model opens new avenues for studying ubiquitination in plants and other species. The implementation is publicly accessible on GitHub at https://github.com/nuinvtnu/LBiPlantUbi .