A study on pharmaceutical text relationship extraction based on heterogeneous graph neural networks.

Effective information extraction of pharmaceutical texts is of great significance for clinical research. The ancient Chinese medicine text has streamlined sentences and complex semantic relationships, and the textual relationships may exist between heterogeneous entities. The current mainstream relationship extraction model does not take into account the associations between entities and relationships when extracting, resulting in insufficient semantic information to form an effective structured representation. In this paper, we propose a heterogeneous graph neural network relationship extraction model adapted to traditional Chinese medicine (TCM) text. First, the given sentence and predefined relationships are embedded by bidirectional encoder representation from transformers (BERT fine-tuned) word embedding as model input. Second, a heterogeneous graph network is constructed to associate words, phrases, and relationship nodes to obtain the hidden layer representation. Then, in the decoding stage, two-stage subject-object entity identification method is adopted, and the identifier adopts a binary classifier to locate the start and end positions of the TCM entities, identifying all the subject-object entities in the sentence, and finally forming the TCM entity relationship group. Through the experiments on the TCM relationship extraction dataset, the results show that the precision value of the heterogeneous graph neural network embedded with BERT is 86.99% and the F1 value reaches 87.40%, which is improved by 8.83% and 10.21% compared with the relationship extraction models CNN, Bert-CNN, and Graph LSTM.

Multiple prescription pattern recognition model based on Siamese network.

Prescription data is an important focus and breakthrough in the study of clinical treatment rules, and the complex multidimensional relationships between Traditional Chinese medicine (TCM) prescription data increase the difficulty of extracting knowledge from clinical data. This paper proposes a complex prescription recognition algorithm (MTCMC) based on the classification and matching of TCM prescriptions with classical prescriptions to identify the classical prescriptions contained in the prescriptions and provide a reference for mining TCM knowledge. The MTCMC algorithm first calculates the importance level of each drug in the complex prescriptions and determines the core prescription combinations of patients through the Analytic Hierarchy Process (AHP) combined with drug dosage. Secondly, a drug attribute tagging strategy was used to quantify the functional features of each drug in the core prescriptions; finally, a Bidirectional Long Short-Term Memory Network (BiLSTM) was used to extract the relational features of the core prescriptions, and a vector representation similarity matrix was constructed in combination with the Siamese network framework to calculate the similarity between the core prescriptions and the classical prescriptions. The experimental results show that the accuracy and F1 score of the prescription matching dataset constructed based on this paper reach 94.45% and 94.34% respectively, which is a significant improvement compared with the models of existing methods.

Dose-effect relationship analysis of TCM based on deep Boltzmann machine and partial least squares.

A dose-effect relationship analysis of traditional Chinese Medicine (TCM) is crucial to the modernization of TCM. However, due to the complex and nonlinear nature of TCM data, such as multicollinearity, it can be challenging to conduct a dose-effect relationship analysis. Partial least squares can be applied to multicollinearity data, but its internally extracted principal components cannot adequately express the nonlinear characteristics of TCM data. To address this issue, this paper proposes an analytical model based on a deep Boltzmann machine (DBM) and partial least squares. The model uses the DBM to extract nonlinear features from the feature space, replaces the components in partial least squares, and performs a multiple linear regression. Ultimately, this model is suitable for analyzing the dose-effect relationship of TCM. The model was evaluated using experimental data from Ma Xing Shi Gan Decoction and datasets from the UCI Machine Learning Repository. The experimental results demonstrate that the prediction accuracy of the model based on the DBM and partial least squares method is on average 10% higher than that of existing methods.

Delayed callose degradation restores the fertility of multiple P/TGMS lines in Arabidopsis.

Photoperiod/temperature-sensitive genic male sterility (P/TGMS) is widely applied for improving crop production. Previous investigations using the reversible male sterile (rvms) mutant showed that slow development is a general mechanism for restoring fertility to P/TGMS lines in Arabidopsis. In this work, we isolated a restorer of rvms-2 (res3), as the male sterility of rvms-2 was rescued by res3. Phenotype analysis and molecular cloning show that a point mutation in UPEX1 l in res3 leads to delayed secretion of callase A6 from the tapetum to the locule and tetrad callose wall degradation. Electrophoretic mobility shift assay and chromatin immunoprecipitation analysis demonstrated that the tapetal transcription factor ABORTED MICROSPORES directly regulates UPEX1 expression, revealing a pathway for tapetum secretory function. Early degradation of the callose wall in the transgenic line eliminated the fertility restoration effect of res3. The fertility of multiple known P/TGMS lines with pollen wall defects was also restored by res3. We propose that the remnant callose wall may broadly compensate for the pollen wall defects of P/TGMS lines by providing protection for pollen formation. A cellular mechanism is proposed to explain how slow development restores the fertility of P/TGMS lines in Arabidopsis.

IMPERFECTIVE EXINE FORMATION (IEF) is required for exine formation and male fertility in Arabidopsis.

KEY MESSAGE: IEF, a novel plasma plasma membrane protein, is important for exine formation in Arabidopsis. Exine, an important part of pollen wall, is crucial for male fertility. The major component of exine is sporopollenin which are synthesized and secreted by tapetum. Although sporopollenin synthesis has been well studied, the transportation of it remains elusive. To understand it, we analyzed the gene expression pattern in tapetal microdissection data, and investigated the potential transporter genes that are putatively regulated by ABORTED MICROSPORES (AMS). Among these genes, we identified IMPERFECTIVE EXINE FORMATION (IEF) that is important for exine formation. Compared to the wild type, ief mutants exhibit severe male sterility and pollen abortion, suggesting IEF is crucial for pollen development and male fertility. Using both scanning and transmission electron microscopes, we showed that exine structure was not well defined in ief mutant. The transient expression of IEF-GFP driven by the 35S promoter indicated that IEF-GFP was localized in plasma membrane. Furthermore, AMS can specifically activate the expression of promoterIEF:LUC in vitro, which suggesting AMS regulates IEF for exine formation. The expression of ATP-BINDING CASSETTE TRANSPORTER G26 (AGCB26) was not affected in ief mutants. In addition, SEM and TEM data showed that the sporopollenin deposition is more defective in abcg26/ief-2 than that of in abcg26, which suggesting that IEF is involved in an independent sporopollenin transportation pathway. This work reveal a novel gene, IEF regulated by AMS that is essential for exine formation.

Bioassay-Guided Isolation of Triterpenoids as α-Glucosidase Inhibitors from Cirsium setosum.

Cirsium setosum (C. setosum) has a potential antihyperglycemic effect, but it is unclear what bioactive components play a key role. According to the α-glucosidase inhibition activity, three new taraxastane-type triterpenoids of 3ß-hydroxy-30-hydroperoxy-20-taraxastene (1), 3ß-hydroxy-22α-methoxy-20-taraxastene (2), and 30-nor-3ß,22α-dihydroxy-20-taraxastene (3), as well as five known taraxastane triterpenoids of 3ß,22-dihydroxy-20-taraxastene (4), 20-taraxastene-3,22-dione (5), 3ß-acetoxy-20-taraxasten-22-one (6), 3ß-hydroxy-20-taraxasten-22-one (7), and 30-nor-3ß-hydroxy-20-taraxastene (8) were obtained from the petroleum ether-soluble portion of the ethanol extract from C. setosum. All chemical structures of the compounds were elucidated by spectroscopic data analysis and compared with literature data. Compounds 4-8 were identified for the first time from this plant, and compounds 1, 2, 4, and 7 exhibited more potent α-glucosidase inhibitory activity-with IC50 values of 18.34 ± 1.27, 26.98 ± 0.89, 17.49 ± 1.42, and 22.67 ± 0.25 µM, respectively-than acarbose did (positive control, IC50 42.52 ± 0.32 µM).