Mining microbe–disease interactions from literature via a transfer learning model

Abstract

Background Interactions of microbes and diseases are of great importance for biomedical research. However, large-scale of microbe–disease interactions are hidden in the biomedical literature. The structured databases for microbe–disease interactions are in limited amounts. In this paper, we aim to construct a large-scale database for microbe–disease interactions automatically. We attained this goal via applying text mining methods based on a deep learning model with a moderate curation cost. We also built a user-friendly web interface that allows researchers to navigate and query required information. Results Firstly, we manually constructed a golden-standard corpus and a sliver-standard corpus (SSC) for microbe–disease interactions for curation. Moreover, we proposed a text mining framework for microbe–disease interaction extraction based on a pretrained model BERE. We applied named entity recognition tools to detect microbe and disease mentions from the free biomedical texts. After that, we fine-tuned the pretrained model BERE to recognize relations between targeted entities, which was originally built for drug–target interactions or drug–drug interactions. The introduction of SSC for model fine-tuning greatly improved detection performance for microbe–disease interactions, with an average reduction in error of approximately 10%. The MDIDB website offers data browsing, custom searching for specific diseases or microbes, and batch downloading. Conclusions Evaluation results demonstrate that our method outperform the baseline model (rule-based PKDE4J) with an average $$F_1$$ F 1 -score of 73.81%. For further validation, we randomly sampled nearly 1000 predicted interactions by our model, and manually checked the correctness of each interaction, which gives a 73% accuracy. The MDIDB webiste is freely avaliable throuth http://dbmdi.com/index/

Large scale text mining for deriving useful insights: A case study focused on microbiome

“…We also did not compare the quality of the networks with existing hand curated manual efforts in this direction, it will interesting future work to perform a data driven comparison to assess the performance. Additionally, it would be interesting to use these embeddings to infer causality or to infer positive or negative correlations between concepts (such as in Wu et al (2021) ). Additionally, in this study we only established the quality of the learnt semantic relationships empirically since our focus was only to establish the potential of text mining algorithms to derive such relationships with no pre-training or domain understanding, a more rigorous analysis of the networks is interesting future work.…”

Hierarchical shared transfer learning for biomedical named entity recognition

“…Where parameter transfer is already commonly used in NLP tasks, it is assumed to share some parameters between source tasks and target tasks, or to share a prior distribution of model hyperparameters [10]. This also enables good accuracy when transferring the original model to the new domain [11,12]. However, there are also problems with negative transfer.…”

Leveraging pre-trained language models for mining microbiome-disease relationships

“…Next is the work by Wu et al[ 24 ] that focuses on a deep-learning strategy for solving this problem. Their approach first involves preparing training datasets.…”

“…Secondly, our approach highlights the relevance of deep learning and transformer models in reducing the requirement for large amounts of training data. In contrast to the study by [ 24 ] or other deep learning models, our method benefits from transfer learning and task-specific fine-tuning. This advantage enables us to achieve excellent results with a smaller amount of training data, making our approach efficient and practical.…”

“…Initially, we explore the potential of pre-trained language models as zero-shot and few-shot learners, without fine-tuning, potentially eliminating the need for training data or specialized retraining and fine-tuning. Subsequently, we fine-tune these models using curated training data, obtained from [ 24 ], to further enhance their capabilities.…”