Scholarlink

Biomarkers play a crucial role in medical care and drug development. Among them, urine has gained attention as a promising source of biomarkers. Previously, we created a database of urinary protein biomarkers by manually curating the literature. In this study, we developed a deep learning–based model called BiomarkerExtractor to enhance text mining efficiency. BiomarkerExtractor consists of a BioBERT-BiLSTM-CRF model for identifying named entities and a BioBERT-FAM model for extracting relationships between biomarkers and diseases. We applied this workflow to identify protein and metabolite biomarkers in urine. The results were reviewed and corrected manually, and the corrected data was used to train the model further. To facilitate the review of biomarker information, we also built a new website with a crowdsourcing platform for user participation and scoring.

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

BiomarkerExtractor: A Deep Learning–Based Model for Text Mining of Urine Biomarkers

  • 章节

  • 发表日期：

    2025年

查看全文
  • Mingze Bai,
  • Xinnan Zhang,
  • Chen Shao

摘要

Biomarkers play a crucial role in medical care and drug development. Among them, urine has gained attention as a promising source of biomarkers. Previously, we created a database of urinary protein biomarkers by manually curating the literature. In this study, we developed a deep learning–based model called BiomarkerExtractor to enhance text mining efficiency. BiomarkerExtractor consists of a BioBERT-BiLSTM-CRF model for identifying named entities and a BioBERT-FAM model for extracting relationships between biomarkers and diseases. We applied this workflow to identify protein and metabolite biomarkers in urine. The results were reviewed and corrected manually, and the corrected data was used to train the model further. To facilitate the review of biomarker information, we also built a new website with a crowdsourcing platform for user participation and scoring.