Contextual word embeddings and the Transformers architecture have reached state-of-the-art results in many natural language processing (NLP) tasks and improved the adaptation of models for multiple domains. Despite the improvement in the reuse and construction of models, few resources are still developed for the Portuguese language, especially in the health domain. Furthermore, the clinical models available for the language are not representative enough for all medical specialties. This work explores deep contextual embedding models for the Portuguese language to support clinical NLP tasks. We transferred learned information from electronic health records of a Brazilian tertiary hospital specialized in cardiology diseases and pre-trained multiple clinical BERT-based models. We evaluated the performance of these models in named entity recognition experiments, fine-tuning them in two annotated corpora containing clinical narratives. Our pre-trained models outperformed previous multilingual and Portuguese BERT-based models for cardiology and multi-specialty environments, reaching the state-of-the-art for analyzed corpora, with 5.5% F1 score improvement in TempClinBr (all entities) and 1.7% in SemClinBr (Disorder entity) corpora. Hence, we demonstrate that data representativeness and a high volume of training data can improve the results for clinical tasks, aligned with results for other languages.
Azopirillum brasilense is a plant-growth promoting nitrogen-fixing bacteria that is used as bio-fertilizer in agriculture. Since nitrogen fixation has a high-energy demand, the reduction of N2 to NH4 + by nitrogenase occurs only under limiting conditions of NH4 + and O2. Moreover, the synthesis and activity of nitrogenase is highly regulated to prevent energy waste. In A. brasilense nitrogenase activity is regulated by the products of draG and draT. The product of the draB gene, located downstream in the draTGB operon, may be involved in the regulation of nitrogenase activity by an, as yet, unknown mechanism. A deep in silico analysis of the product of draB was undertaken aiming at suggesting its possible function and involvement with DraT and DraG in the regulation of nitrogenase activity in A. brasilense. In this work, we present a new artificial intelligence strategy for protein classification, named ProClaT. The features used by the pattern recognition model were derived from the primary structure of the DraB homologous proteins, calculated by a ProClaT internal algorithm. ProClaT was applied to this case study and the results revealed that the A. brasilense draB gene codes for a protein highly similar to the nitrogenase associated NifO protein of Azotobacter vinelandii. This tool allowed the reclassification of DraB/NifO homologous proteins, hypothetical, conserved hypothetical and those annotated as putative arsenate reductase, ArsC, as NifO-like. An analysis of co-occurrence of draB, draT, draG and of other nif genes was performed, suggesting the involvement of draB (nifO) in nitrogen fixation, however, without the definition of a specific function.
20 January 2026 in Journal of the Brazilian Computer Society
Named Entity Recognition (NER) is essential in Natural Language Processing (NLP) for extracting pertinent information from unstructured data. Traditional NER approaches assume continuous and non-overlapping entities, which can be limiting in real-world scenarios. This research introduces BioNestedNER, a hybrid method for nested, discontinuous, and multi-type entity recognition, with a focus on clinical and biomedical domains. Our approach employs a language model (encoder-only Transformer-based model) using a machine reading comprehension strategy, treating NER as a question-answering-like task. A Conditional Random Field also addresses multi-label sequence labeling for handling nested entities as multi-type entities. Evaluation in Portuguese demonstrated state-of-the-art performance in micro F1-Scores across two clinical corpora. In NestedClinBr, featuring nested and discontinuous entities, our method achieved an F1-Score of 0.863, surpassing the second-place result by 2.1%. In SemClinBr, with multi-type entities, an F1-Score of 0.782 was achieved, surpassing the second-place result by 11.5%. This paper also presents a new clinical corpus in Brazilian Portuguese annotated with nested and discontinuous entities, offering a valuable resource for developing and evaluating models handling these complex entities. In conclusion, BioNestedNER presents an adaptable and effective NER solution for nested, discontinuous, and multi-type entities, with the potential to benefit various clinical applications.