BioBERT-based SNP-trait Associations Extraction from Biomedical Literature

A vast amount of information available in scientific literature presents a valuable opportunity to develop text-mining techniques for extracting biomedical relationships. One crucial area of interest involves analyzing the connection between singular nucleotide polymorphism (SNP) and traits. In this study, we introduce BioBERT-GRU to identify the associations between SNP and traits. Through evaluating our approach using the SNPPhenA dataset, we have determined that this novel method outperforms previous machine learning and deep learning techniques that have been used for a similar aim. BioBERT-GRU achieved a significant result, demonstrating a precision, recall, and F1-score equal to 88.3%, 88.2%, 88.1%, respectively.