Beyond MeSH: Fine-Grained Semantic Indexing of Biomedical Literature Based on Weak Supervision

Abstract

In this work, we propose a method for the automated refinement of subject annotations in biomedical literature at the level of concepts. Semantic indexing and search of biomedical articles in MEDLINE/PubMed are based on semantic subject annotations with MeSH descriptors that may correspond to several related but distinct biomedical concepts. Such semantic annotations do not adhere to the level of detail available in the domain knowledge and may not be sufficient to fulfil the information needs of experts in the domain. To this end, we propose a new method that uses weak supervision to train a concept annotator on the literature available for a particular disease. We test this method on the MeSH descriptors for two diseases: Alzheimer's Disease and Duchenne Muscular Dystrophy. The results indicate that concept-occurrence is a strong heuristic for automated subject annotation refinement and its use as weak supervision can lead to improved concept-level annotations. The fine-grained semantic annotations can enable more precise literature retrieval, sustain the semantic integration of subject annotations with other domain resources and ease the maintenance of consistent subject annotations, as new more detailed entries are added in the MeSH thesaurus over time.

Large-scale fine-grained semantic indexing of biomedical literature based on weakly-supervised deep learning

“…More recently, we proposed the Beyond MeSH method [46,2] for the automated refinement of MeSH topic annotations at the level of MeSH concepts.…”

“…In particular, we focus on the year y, when the promotion of c to d c took place, as shown in Figure 4. All the articles annotated with d 16 prior to year y are gathered for the development of a weakly-supervised development dataset for indexing with c. The fact that c used to be a subordinate concept of d in this period, ensures that any article relevant to c will be included in this set of articles annotated with d. To generate the weak labels we use the heuristic introduced in Beyond MeSH [46,2], which relies on the occurrence of concept c in the title or abstract of the article, as identified by the MetaMap tool [47].…”