RESUMO
We report a combined manual annotation and deep-learning natural language processing study to make accurate entity extraction in hereditary disease related biomedical literature. A total of 400 full articles were manually annotated based on published guidelines by experienced genetic interpreters at Beijing Genomics Institute (BGI). The performance of our manual annotations was assessed by comparing our re-annotated results with those publicly available. The overall Jaccard index was calculated to be 0.866 for the four entity types-gene, variant, disease and species. Both a BERT-based large name entity recognition (NER) model and a DistilBERT-based simplified NER model were trained, validated and tested, respectively. Due to the limited manually annotated corpus, Such NER models were fine-tuned with two phases. The F1-scores of BERT-based NER for gene, variant, disease and species are 97.28%, 93.52%, 92.54% and 95.76%, respectively, while those of DistilBERT-based NER are 95.14%, 86.26%, 91.37% and 89.92%, respectively. Most importantly, the entity type of variant has been extracted by a large language model for the first time and a comparable F1-score with the state-of-the-art variant extraction model tmVar has been achieved.
RESUMO
CORRECTION: Following publication of the original article [1], the authors reported that there was a mistake in the presentation of their funding information. The sentence "This study was supported by the National Natural Science Foundation of China (31,100,268 to Peng Chen, 31,270,658 to Bo Zheng);" should instead read "This study was supported by the National Natural Science Foundation of China (31100268 to Peng Chen, 31370604 to Bo Zheng);".
RESUMO
BACKGROUND: Modification of nucleosides on transfer RNA (tRNA) is important either for correct mRNA decoding process or for tRNA structural stabilization. Nucleoside methylations catalyzed by MTase (methyltransferase) are the most common type among all tRNA nucleoside modifications. Although tRNA modified nucleosides and modification enzymes have been extensively studied in prokaryotic systems, similar research remains preliminary in higher plants, especially in crop species, such as rice (Oryza sativa). Rice is a monocot model plant as well as an important cereal crop, and stress tolerance and yield are of great importance for rice breeding. RESULTS: In this study, we investigated how the composition and abundance of tRNA modified nucleosides could change in response to drought, salt and cold stress, as well as in different tissues during the whole growth season in two model plants-O. sativa and Arabidopsis thaliana. Twenty two and 20 MTase candidate genes were identified in rice and Arabidopsis, respectively, by protein sequence homology and conserved domain analysis. Four methylated nucleosides, Am, Cm, m1A and m7G, were found to be very important in stress response both in rice and Arabidopsis. Additionally, three nucleosides,Gm, m5U and m5C, were involved in plant development. Hierarchical clustering analysis revealed consistency on Am, Cm, m1A and m7G MTase candidate genes, and the abundance of the corresponding nucleoside under stress conditions. The same is true for Gm, m5U and m5C modifications and corresponding methylation genes in different tissues during different developmental stages. CONCLUSIONS: We identified candidate genes for various tRNA modified nucleosides in rice and Arabidopsis, especially on MTases for methylated nucleosides. Based on bioinformatics analysis, nucleoside abundance assessments and gene expression profiling, we propose four methylated nucleosides (Am, Cm, m1A and m7G) that are critical for stress response in rice and Arabidopsis, and three methylated nucleosides (Gm, m5U and m5C) that might be important during development.
