Annotating publicly-available samples and studies using interpretable modeling of unstructured metadata

Abstract

Reusing massive collections of publicly available biomedical data can significantly impact knowledge discovery. However, these public samples and studies are typically described using unstructured plain text, hindering the findability and further reuse of the data. To combat this problem, we proposetxt2onto 2.0, a general-purpose method based on natural language processing and machine learning for annotating biomedical unstructured metadata to controlled vocabularies of diseases and tissues. Compared to the previous version (txt2onto 1.0), which uses numerical embeddings as features, this new version uses words as features, resulting in improved interpretability and performance, especially when few positive training instances are available.Txt2onto 2.0uses embeddings from a large language model during prediction to deal with unseen-yet-relevant words in the input text and to highlight biomedical concepts in the input text that are related to each disease and tissue term being predicted, thereby explaining the basis of every annotation. We demonstrate the generalizability oftxt2onto 2.0by accurately predicting disease annotations for studies from independent datasets, using proteomics and clinical trials as examples. Overall, our approach can annotate biomedical text regardless of experimental types or sources. Code, data, and trained models are available athttps://github.com/krishnanlab/txt2onto2.0.