iNucRes-ASSH: Identifying nucleic acid-binding residues in proteins by using self-attention-based structure-sequence hybrid neural network.

Interaction between proteins and nucleic acids is crucial to many cellular activities. Accurately detecting nucleic acid-binding residues (NABRs) in proteins can help researchers better understand the interaction mechanism between proteins and nucleic acids. Structure-based methods can generally make more accurate predictions than sequence-based methods. However, the existing structure-based methods are sensitive to protein conformational changes, causing limited generalizability. More effective and robust approaches should be further explored. In this study, we propose iNucRes-ASSH to identify nucleic acid-binding residues with a self-attention-based structure-sequence hybrid neural network. It improves the generalizability and robustness of NABR prediction from two levels: residue representation and prediction model. Experimental results show that iNucRes-ASSH can predict the nucleic acid-binding residues even when the experimentally validated structures are unavailable and outperforms five competing methods on a recent benchmark dataset and a widely used test dataset.

The effect of sevoflurane exposure on cell-type-specific changes in the prefrontal cortex in young mice.

Sevoflurane, the predominant pediatric anesthetic, has been linked to neurotoxicity in young mice, although the underlying mechanisms remain unclear. This study focuses on investigating the impact of neonatal sevoflurane exposure on cell-type-specific alterations in the prefrontal cortex (PFC) of young mice. Neonatal mice were subjected to either control treatment (60% oxygen balanced with nitrogen) or sevoflurane anesthesia (3% sevoflurane in 60% oxygen balanced with nitrogen) for 2 hours on postnatal days (PNDs) 6, 8, and 10. Behavioral tests and single-nucleus RNA sequencing (snRNA-seq) of the PFC were conducted from PNDs 31 to 37. Mechanistic exploration included clustering analysis, identification of differentially expressed genes (DEGs), enrichment analyses, single-cell trajectory analysis, and genome-wide association studies (GWAS). Sevoflurane anesthesia resulted in sociability and cognition impairments in mice. Novel specific marker genes identified 8 distinct cell types in the PFC. Most DEGs between the control and sevoflurane groups were unique to specific cell types. Re-defining 15 glutamatergic neuron subclusters based on layer identity revealed their altered expression profiles. Notably, sevoflurane disrupted the trajectory from oligodendrocyte precursor cells (OPCs) to oligodendrocytes (OLs). Validation of disease-relevant candidate genes across the main cell types demonstrated their association with social dysfunction and working memory impairment. Behavioral results and snRNA-seq collectively elucidated the cellular atlas in the PFC of young male mice, providing a foundation for further mechanistic studies on developmental neurotoxicity induced by anesthesia.

NCBRPred: predicting nucleic acid binding residues in proteins based on multilabel learning.

The interactions between proteins and nucleic acid sequences play many important roles in gene expression and some cellular activities. Accurate prediction of the nucleic acid binding residues in proteins will facilitate the research of the protein functions, gene expression, drug design, etc. In this regard, several computational methods have been proposed to predict the nucleic acid binding residues in proteins. However, these methods cannot satisfactorily measure the global interactions among the residues along protein. Furthermore, these methods are suffering cross-prediction problem, new strategies should be explored to solve this problem. In this study, a new computational method called NCBRPred was proposed to predict the nucleic acid binding residues based on the multilabel sequence labeling model. NCBRPred used the bidirectional Gated Recurrent Units (BiGRUs) to capture the global interactions among the residues, and treats this task as a multilabel learning task. Experimental results on three widely used benchmark datasets and an independent dataset showed that NCBRPred achieved higher predictive results with lower cross-prediction, outperforming 10 existing state-of-the-art predictors. The web-server and a stand-alone package of NCBRPred are freely available at http://bliulab.net/NCBRPred. It is anticipated that NCBRPred will become a very useful tool for identifying nucleic acid binding residues.

PreRBP-TL: prediction of species-specific RNA-binding proteins based on transfer learning.

MOTIVATION: RNA-binding proteins (RBPs) play crucial roles in post-transcriptional regulation. Accurate identification of RBPs helps to understand gene expression, regulation, etc. In recent years, some computational methods were proposed to identify RBPs. However, these methods fail to accurately identify RBPs from some specific species with limited data, such as bacteria. RESULTS: In this study, we introduce a computational method called PreRBP-TL for identifying species-specific RBPs based on transfer learning. The weights of the prediction model were initialized by pretraining with the large general RBP dataset and then fine-tuned with the small species-specific RPB dataset by using transfer learning. The experimental results show that the PreRBP-TL achieves better performance for identifying the species-specific RBPs from Human, Arabidopsis, Escherichia coli and Salmonella, outperforming eight state-of-the-art computational methods. It is anticipated PreRBP-TL will become a useful method for identifying RBPs. AVAILABILITY AND IMPLEMENTATION: For the convenience of researchers to identify RBPs, the web server of PreRBP-TL was established, freely available at http://bliulab.net/PreRBP-TL. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Single-nucleus Atlas of Sevoflurane-induced Hippocampal Cell Type- and Sex-specific Effects during Development in Mice.

BACKGROUND: Multiple neonatal exposures to sevoflurane induce neurocognitive dysfunctions in rodents. The lack of cell type-specific information after sevoflurane exposure limits the mechanistic understanding of these effects. In this study, the authors tested the hypothesis that sevoflurane exposures alter the atlas of hippocampal cell clusters and have neuronal and nonneuronal cell type-specific effects in mice of both sexes. METHODS: Neonatal mice were exposed to 3% sevoflurane for 2 h at postnatal days 6, 8, and 10 and analyzed for the exposure effects at postnatal day 37. Single-nucleus RNA sequencing was performed in the hippocampus followed by in situ hybridization to validate the results of RNA sequencing. The Morris Water Maze test was performed to test neurocognitive function. RESULTS: The authors found sex-specific distribution of hippocampal cell types in control mice alongside cell type- and sex-specific effects of sevoflurane exposure on distinct hippocampal cell populations. There were important changes in male but not in female mice after sevoflurane exposure regarding the proportions of cornu ammonis 1 neurons (control vs. sevoflurane, males: 79.9% vs. 32.3%; females: 27.3% vs. 24.3%), dentate gyrus (males: 4.2% vs. 23.4%; females: 36.2% vs. 35.8%), and oligodendrocytes (males: 0.6% vs. 6.9%; females: 5.9% vs. 7.8%). In male but not in female mice, sevoflurane altered the number of significantly enriched ligand-receptor pairs in the cornu ammonis 1, cornu ammonis 3, and dente gyrus trisynaptic circuit (control vs. sevoflurane, cornu ammonis 1-cornu ammonis 3: 18 vs. 42 in males and 15 vs. 21 in females; cornu ammonis 1-dentate gyrus: 21 vs. 35 in males and 12 vs. 20 in females; cornu ammonis 3-dentate gyrus: 25 vs. 45 in males and 17 vs. 20 in females), interfered with dentate gyrus granule cell neurogenesis, hampered microglia differentiation, and decreased cornu ammonis 1 pyramidal cell diversity. Oligodendrocyte differentiation was specifically altered in females with increased expressions of Mbp and Mag. In situ hybridization validated the increased expression of common differentially expressed genes. CONCLUSIONS: This single-nucleus RNA sequencing study reveals the hippocampal atlas of mice, providing a comprehensive resource for the neuronal and nonneuronal cell type- and sex-specific effects of sevoflurane during development.

Outdoor Health Risk of Atmospheric Particulate Matter at Night in Xi'an, Northwestern China.

The deterioration of air quality via anthropogenic activities during the night period has been deemed a serious concern among the scientific community. Thereby, we explored the outdoor particulate matter (PM) concentration and the contributions from various sources during the day and night in winter and spring 2021 in a megacity, northwestern China. The results revealed that the changes in chemical compositions of PM and sources (motor vehicles, industrial emissions, coal combustion) at night lead to substantial PM toxicity, oxidative potential (OP), and OP/PM per unit mass, indicating high oxidative toxicity and exposure risk at nighttime. Furthermore, higher environmentally persistent free radical (EPFR) concentration and its significant correlation with OP were observed, suggesting that EPFRs cause reactive oxygen species (ROS) formation. Moreover, the noncarcinogenic and carcinogenic risks were systematically explained and spatialized to children and adults, highlighting intensified hotspots to epidemiological researchers. This better understanding of day-night-based PM formation pathways and their hazardous impact will assist to guide measures to diminish the toxicity of PM and reduce the disease led by air pollution.

Biomedical relation extraction via knowledge-enhanced reading comprehension.

BACKGROUND: In biomedical research, chemical and disease relation extraction from unstructured biomedical literature is an essential task. Effective context understanding and knowledge integration are two main research problems in this task. Most work of relation extraction focuses on classification for entity mention pairs. Inspired by the effectiveness of machine reading comprehension (RC) in the respect of context understanding, solving biomedical relation extraction with the RC framework at both intra-sentential and inter-sentential levels is a new topic worthy to be explored. Except for the unstructured biomedical text, many structured knowledge bases (KBs) provide valuable guidance for biomedical relation extraction. Utilizing knowledge in the RC framework is also worthy to be investigated. We propose a knowledge-enhanced reading comprehension (KRC) framework to leverage reading comprehension and prior knowledge for biomedical relation extraction. First, we generate questions for each relation, which reformulates the relation extraction task to a question answering task. Second, based on the RC framework, we integrate knowledge representation through an efficient knowledge-enhanced attention interaction mechanism to guide the biomedical relation extraction. RESULTS: The proposed model was evaluated on the BioCreative V CDR dataset and CHR dataset. Experiments show that our model achieved a competitive document-level F1 of 71.18% and 93.3%, respectively, compared with other methods. CONCLUSION: Result analysis reveals that open-domain reading comprehension data and knowledge representation can help improve biomedical relation extraction in our proposed KRC framework. Our work can encourage more research on bridging reading comprehension and biomedical relation extraction and promote the biomedical relation extraction.

Significant Promotion of Light Absorption Ability and Formation of Triplet Organics and Reactive Oxygen Species in Atmospheric HULIS by Fe(III) Ions.

Metal ions are key components in atmosphere that potentially affect the optical properties and photochemical reactivity of atmospheric humic-like substances (HULIS), while this mechanism is still unclear. In this study, we demonstrated that atmospheric HULIS coupled with Fe3+, Cu2+, Zn2+, and Al3+ exhibited distinct optical properties and reactive intermediates from that of HULIS utilizing three-dimensional fluorescence spectroscopy and electron paramagnetic resonance spectroscopy. The HULIS components showed light absorption that increased by 56% for the HULIS-Fe3+ system, fluorescence blue shift, and fluorescence quenching, showing a certain dose-effect relationship. These are mainly attributed to the fact that the highly oxidative HULIS chromophores have a stronger complexing ability with Fe3+ ions than the other metal ions. In addition, triplet organics (promoting ratio: 53%) and reactive oxygen species (promoting ratio: 82.6%) in the HULIS-Fe3+ system showed obvious generation promotion. Therefore, the main assumption of the photochemical mechanisms of atmospheric HULIS in the HULIS-Fe3+ system is that Fe3+ ions can form 3HULIS*-Fe3+ complexation with photoexcited 3HULIS* and then transition to the ground state through energy transfer, electron transfer, or nonradiative transition, accompanied by the formation of singlet oxygen and hydroxyl radicals. Our results provide references for evaluating the radiative forcing and aging effect of metal ions on atmospheric aerosols.

Connecting the Light Absorption of Atmospheric Organic Aerosols with Oxidation State and Polarity.

The light-absorbing organic aerosol (OA) constitutes an important fraction of absorbing components, counteracting major cooling effect of aerosols to climate. The mechanisms in linking the complex and changeable chemistry of OA with its absorbing properties remain to be elucidated. Here, by using solvent extraction, ambient OA from an urban environment was fractionated according to polarity, which was further nebulized and online characterized with compositions and absorbing properties. Water extracted high-polar compounds with a significantly higher oxygen to carbon ratio (O/C) than methanol extracts. A transition O/C of about 0.6 was found, below and above which the enhancement and reduction of OA absorptivity were observed with increasing O/C, occurring on the less polar and high polar compounds, respectively. In particular, the co-increase of nitrogen and oxygen elements suggests the important role of nitrogen-containing functional groups in enhancing the absorptivity of the less polar compounds (e.g., forming nitrogen-containing aromatics), while further oxidation (O/C > 0.6) on high-polar compounds likely led to fragmentation and bleaching chromophores. The results here may reconcile the previous observations about darkening or whitening chromophores of brown carbon, and the parametrization of O/C has the potential to link the changing chemistry of OA with its polarity and absorbing properties.

Leveraging Multi-source knowledge for Chinese clinical named entity recognition via relational graph convolutional network.

OBJECTIVE: External knowledge, such as lexicon of words in Chinese and domain knowledge graph (KG) of concepts, has been recently adopted to improve the performance of machine learning methods for named entity recognition (NER) as it can provide additional information beyond context. However, most existing studies only consider knowledge from one source (i.e., either lexicon or knowledge graph) in different ways and consider lexicon words or KG concepts independently with their boundaries. In this paper, we focus on leveraging multi-source knowledge in a unified manner where lexicon words or KG concepts are well combined with their boundaries for Chinese Clinical NER (CNER). MATERIAL AND METHODS: We propose a novel method based on relational graph convolutional network (RGCN), called MKRGCN, to utilize multi-source knowledge in a unified manner for CNER. For any sentence, a relational graph based on words or concepts in each knowledge source is constructed, where lexicon words or KG concepts appearing in the sentence are linked to the containing tokens with the boundary information of the lexicon words or KG concepts. RGCN is used to model all relational graphs constructed from multi-source knowledge, and the representations of tokens from multi-source knowledge are integrated into the context representations of tokens via an attention mechanism. Based on the knowledge-enhanced representations of tokens, we deploy a conditional random field (CRF) layer for named entity label prediction. In this study, a lexicon of words and a medical knowledge graph are used as knowledge sources for Chinese CNER. RESULTS: Our proposed method achieves the best performance on CCKS2017 and CCKS2018 in Chinese with F1-scores of 91.88% and 89.91%, respectively, significantly outperforming existing methods. The extended experiments on NCBI-Disease and BC2GM in English also prove the effectiveness of our method when only considering one knowledge source via RGCN. CONCLUSION: The MKRGCN model can integrate knowledge from the external lexicon and knowledge graph effectively for Chinese CNER and has the potential to be applied to English NER.

Oxidation potential and coupling effects of the fractionated components in airborne fine particulate matter.

Fine particulate matter (PM2.5) can induce the generation of reactive oxygen species (ROS) and damage human tissues. Fully understanding the generation mechanism of oxidative toxicity of PM is challenging due to the extremely complex composition. Classification methods may be helpful in understanding the ROS production mechanisms of complex PM. This study used a solvent extraction and solid phase extraction methods to separate five different components from PM2.5 includes non-extractable components that have rarely been studied before, and discussed the coupling effect and heterogeneous characteristics of oxidation activity they produced. It is found that the water-soluble component contribute about half of the PM oxidation activity, and metal ions probably contribute most of the oxidation activity. Experimental results show that oxygen molecules is the main precursor of ROS production, which depends on whether the aerosol component has catalytic conversion ability. After mixing humic-like substance (HULIS) and hydrophilic water-soluble (HP-WSM) PM, the oxidation activity increased, it is most likely to be a synergistic effect between HULIS and metal ions is dominant, but limited contribution to oxidation activity. It turns out that the non-extractable and water-insoluble components have higher oxidation activity than the water-soluble components, and the two components exhibited a more durable ability to produce 1O2. The reaction of soluble components to produce ROS is homogeneous, but it is obviously heterogeneous for these insoluble components. This study suggests that future attention should be paid to the oxidative toxicity of the non-extractable component, and that single PM component or compound cannot simply be studied independently.

Pollution characteristics and sources of environmentally persistent free radicals and oxidation potential in fine particulate matter related to city lockdown (CLD) in Xi'an, China.

The impact of COVID-19 control on air quality have been prevalent for the past two years, however few studies have explored the toxicity of atmospheric particulate matter during the epidemic control. Therefore, this research highlights the characteristics and sources of oxidative potential (OP) and the new health risk substances environmentally persistent free radicals (EPFRs) in comparison to city lockdown (CLD) with early days of 2019-2020. Daily particulate matter (PM2.5) samples were collected from January 14 to February 3, 2020, with the same period during 2019 in Xi'an city. The results indicated that the average concentration of PM2.5 decreased by 48% during CLD. Concentrations of other air pollutants and components, such as PM10, NO2, SO2, WSIs, OC and EC were also decreased by 22%, 19%, 2%, 17%, 6%, and 4% respectively during the CLD, compared to the same period in 2019. Whereas only O3 increased by 30% during CLD. The concentrations of EPFRs in PM2.5 was considerably lower than in 2019, which decreased by 12% during CLD. However, the OP level was increased slightly during CLD. Moreover, both EPFRs/PM and DTTv/PM did not decrease or even increase significantly, manifesting that the toxicity of particulate matter has not been reduced by more gains during the CLD. Based on PMF analysis, during the epidemic period, the contribution of traffic emission is significantly reduced, while EPFRs and DTTv increased, which consist of significant O3 and secondary aerosols. This research leads to able future research on human health effect of EPFRs and oxidative potential and can be also used to formulate the majors to control EPFRs and OP emissions, suggest the need for further studies on the secondary processing of EPFRs and OP during the lockdown period in Xi'an. .The COVID-19 lockdown had a significant impact on both social and economic aspects. The city lockdown, however, had a positive impact on the environment and improved air quality, however, no significant health benefits were observed in Xi'an, China.

Remote Ischemic Preconditioning Reduces Acute Kidney Injury After Cardiac Surgery: A Systematic Review and Meta-analysis of Randomized Controlled Trials.

BACKGROUND: Results from previous studies evaluating the effects of remote ischemic preconditioning (RIPC) on morbidity and mortality after cardiac surgery are inconsistent. This meta-analysis of randomized controlled trials (RCTs) aims to determine whether RIPC improves cardiac and renal outcomes in adults undergoing cardiac surgery. METHODS: PubMed, EMBASE, and Cochrane Library were comprehensively searched to identify RCTs comparing RIPC with control in cardiac surgery. The coprimary outcomes were the incidence of postoperative myocardial infarction (MI) and the incidence of postoperative acute kidney injury (AKI). Meta-analyses were performed using a random-effect model. Subgroup analyses were conducted according to volatile only anesthesia versus propofol anesthesia with or without volatiles, high-risk patients versus non-high-risk patients, and Acute Kidney Injury Network (AKIN) or Kidney Disease Improving Global Outcomes (KDIGO) criteria versus other criteria for AKI diagnosis. RESULTS: A total of 79 RCTs with 10,814 patients were included. While the incidence of postoperative MI did not differ between the RIPC and control groups (8.2% vs 9.7%; risk ratio [RR] = 0.87, 95% confidence interval [CI], 0.76-1.01, P = .07, I2 = 0%), RIPC significantly reduced the incidence of postoperative AKI (22% vs 24.4%; RR = 0.86, 95% CI, 0.77-0.97, P = .01, I2 = 34%). The subgroup analyses showed that RIPC was associated with a reduced incidence of MI in non-high-risk patients, and that RIPC was associated with a reduced incidence of AKI in volatile only anesthesia, in non-high-risk patients, and in the studies using AKIN or KDIGO criteria for AKI diagnosis. CONCLUSIONS: This meta-analysis demonstrates that RIPC reduces the incidence of AKI after cardiac surgery. This renoprotective effect of RIPC is mainly evident during volatile only anesthesia, in non-high-risk patients, and when AKIN or KDIGO criteria used for AKI diagnosis.

Water-soluble matter in PM2.5 in a coastal city over China: Chemical components, optical properties, and source analysis.

Although marine and terrestrial emissions simultaneously affect the formation of atmospheric fine particles in coastal areas, knowledge on the optical properties and sources of water-soluble matter in these areas is still scarce. In this work, taking Qingdao, China as a typical coastal location, the chemical composition of PM2.5 during winter 2019 was analyzed. Excitation-emission matrix fluorescence spectroscopy was combined with parallel factor analysis model to explain the components of water-soluble atmospheric chromophores of PM2.5. Our analysis indicated that NO3-, NH4+ and SO42- ions accounted for 86.80% of the total ion mass, dominated by NO3-. The ratio of [NO3-]/[SO42-] was up to 2.42 ± 0.84, suggesting that mobile sources play an important role in local pollutants emission. The result of positive correlation between Abs365 with K+ suggests that biomass burning is an important source of water-soluble organic compounds (WSOC). Six types of fluorophores (C1-C6), all humic-like substances, were identified in WSOC. Humification index, biological index and fluorescence index in winter were 1.66 ± 0.34, 0.51 ± 0.44 and 1.09 ± 0.78, respectively, indicating that WSOC in Qingdao were mainly terrestrial organic matters. Overall, although the study area is close to the ocean, the contribution of terrestrial sources to PM2.5, especially vehicle exhaust and coal combustion, is still much higher than that of marine sources. Our study provides a more comprehensive understanding of chemical and optical properties of WSOC based on PM2.5 in coastal areas, and may provide ground for improving local air quality.

Improving deep learning method for biomedical named entity recognition by using entity definition information.

BACKGROUND: Biomedical named entity recognition (NER) is a fundamental task of biomedical text mining that finds the boundaries of entity mentions in biomedical text and determines their entity type. To accelerate the development of biomedical NER techniques in Spanish, the PharmaCoNER organizers launched a competition to recognize pharmacological substances, compounds, and proteins. Biomedical NER is usually recognized as a sequence labeling task, and almost all state-of-the-art sequence labeling methods ignore the meaning of different entity types. In this paper, we investigate some methods to introduce the meaning of entity types in deep learning methods for biomedical NER and apply them to the PharmaCoNER 2019 challenge. The meaning of each entity type is represented by its definition information. MATERIAL AND METHOD: We investigate how to use entity definition information in the following two methods: (1) SQuad-style machine reading comprehension (MRC) methods that treat entity definition information as query and biomedical text as context and predict answer spans as entities. (2) Span-level one-pass (SOne) methods that predict entity spans of one type by one type and introduce entity type meaning, which is represented by entity definition information. All models are trained and tested on the PharmaCoNER 2019 corpus, and their performance is evaluated by strict micro-average precision, recall, and F1-score. RESULTS: Entity definition information brings improvements to both SQuad-style MRC and SOne methods by about 0.003 in micro-averaged F1-score. The SQuad-style MRC model using entity definition information as query achieves the best performance with a micro-averaged precision of 0.9225, a recall of 0.9050, and an F1-score of 0.9137, respectively. It outperforms the best model of the PharmaCoNER 2019 challenge by 0.0032 in F1-score. Compared with the state-of-the-art model without using manually-crafted features, our model obtains a 1% improvement in F1-score, which is significant. These results indicate that entity definition information is useful for deep learning methods on biomedical NER. CONCLUSION: Our entity definition information enhanced models achieve the state-of-the-art micro-average F1 score of 0.9137, which implies that entity definition information has a positive impact on biomedical NER detection. In the future, we will explore more entity definition information from knowledge graph.

Maternal anesthesia with sevoflurane during the mid-gestation induces social interaction deficits in offspring C57BL/6 mice.

Sevoflurane anesthesia in pregnant mice could induce neurotoxicity in the developing brain and disturb learning and memory in the offspring mice. Whether it could impair social behaviors in the offspring mice is uncertain. Therefore, we assessed the neurobehavioral effect of in-utero exposure to sevoflurane on social interaction behaviors in C57BL/6 mice. The pregnant mice were anesthetized with 2.5% sevoflurane in 100% oxygen for 2 h, and their offspring mice were tested in three-chambered social paradigm, which includes three 10-min sessions of habituation, sociability, and preference for social novelty. At the juvenile age, the offspring mice showed abnormal sociability, as proved by not taking more time sniffing at the stranger 1 mouse compared with the empty enclosure (108.5 ± 25.4 vs. 108.2 ± 44.0 s, P = 0.9876). Meanwhile, these mice showed impaired preference for social novelty, as evidenced by not taking more time sniffing at the stranger 2 compared with the stranger 1 mouse (92.1 ± 52.2 vs. 126.7 ± 50.8 s, P = 0.1502). At the early adulthood, the offspring mice retrieved the normal sociability (145.6 ± 33.2 vs. 76.0 ± 31.8 s, P = 0.0001), but remained the impaired preference for social novelty (100.6 ± 29.3 vs. 118.0 ± 47.9 s, P = 0.3269). Collectively, these results suggested maternal anesthesia with sevoflurane could induce social interaction deficits in their offspring mice. Although the disturbance of sociability could be recoverable, the impairment of preference for social novelty could be long-lasting.

Distinctive Sources Govern Organic Aerosol Fractions with Different Degrees of Oxygenation in the Urban Atmosphere.

Understanding how the sources of an atmospheric organic aerosol (OA) govern its burden is crucial for assessing its impact on the environment and adopting proper control strategies. In this study, the sources of OA over Beijing were assessed year-around based on the combination of two separation approaches for OA, one from chemical fractionation into the high-polarity fraction of water-soluble organic matter (HP-WSOM), humic-like substances (HULIS), and water-insoluble organic matter (WISOM), and the other from statistical grouping using positive matrix factorization (PMF) of high-resolution aerosol mass spectra. Among the three OA fractions, HP-WSOM has the highest O/C ratio (1.36), followed by HULIS (0.56) and WISOM (0.17). The major sources of different OA fractions were distinct: HP-WSOM was dominated by more oxidized oxygenated OA (96%); HULIS by cooking-like OA (40%), less oxidized oxygenated OA (27%), and biomass burning OA (21%); and WISOM by fossil fuel OA (77%). In addition, our results provide evidence that mass spectral-based PMF factors are associated with specific substructures in molecules. These structures are further discussed in the context of the FT-IR results. This study presents an overall relationship of OA groups monitored by chemical and statistical approaches for the first time, providing insights for future source apportionment studies.

Drug knowledge discovery via multi-task learning and pre-trained models.

BACKGROUND: Drug repurposing is to find new indications of approved drugs, which is essential for investigating new uses for approved or investigational drug efficiency. The active gene annotation corpus (named AGAC) is annotated by human experts, which was developed to support knowledge discovery for drug repurposing. The AGAC track of the BioNLP Open Shared Tasks using this corpus is organized by EMNLP-BioNLP 2019, where the "Selective annotation" attribution makes AGAC track more challenging than other traditional sequence labeling tasks. In this work, we show our methods for trigger word detection (Task 1) and its thematic role identification (Task 2) in the AGAC track. As a step forward to drug repurposing research, our work can also be applied to large-scale automatic extraction of medical text knowledge. METHODS: To meet the challenges of the two tasks, we consider Task 1 as the medical name entity recognition (NER), which cultivates molecular phenomena related to gene mutation. And we regard Task 2 as a relation extraction task, which captures the thematic roles between entities. In this work, we exploit pre-trained biomedical language representation models (e.g., BioBERT) in the information extraction pipeline for mutation-disease knowledge collection from PubMed. Moreover, we design the fine-tuning framework by using a multi-task learning technique and extra features. We further investigate different approaches to consolidate and transfer the knowledge from varying sources and illustrate the performance of our model on the AGAC corpus. Our approach is based on fine-tuned BERT, BioBERT, NCBI BERT, and ClinicalBERT using multi-task learning. Further experiments show the effectiveness of knowledge transformation and the ensemble integration of models of two tasks. We conduct a performance comparison of various algorithms. We also do an ablation study on the development set of Task 1 to examine the effectiveness of each component of our method. RESULTS: Compared with competitor methods, our model obtained the highest Precision (0.63), Recall (0.56), and F-score value (0.60) in Task 1, which ranks first place. It outperformed the baseline method provided by the organizers by 0.10 in F-score. The model shared the same encoding layers for the named entity recognition and relation extraction parts. And we obtained a second high F-score (0.25) in Task 2 with a simple but effective framework. CONCLUSIONS: Experimental results on the benchmark annotation of genes with active mutation-centric function changes corpus show that integrating pre-trained biomedical language representation models (i.e., BERT, NCBI BERT, ClinicalBERT, BioBERT) into a pipe of information extraction methods with multi-task learning can improve the ability to collect mutation-disease knowledge from PubMed.

Document-level medical relation extraction via edge-oriented graph neural network based on document structure and external knowledge.

OBJECTIVE: Relation extraction (RE) is a fundamental task of natural language processing, which always draws plenty of attention from researchers, especially RE at the document-level. We aim to explore an effective novel method for document-level medical relation extraction. METHODS: We propose a novel edge-oriented graph neural network based on document structure and external knowledge for document-level medical RE, called SKEoG. This network has the ability to take full advantage of document structure and external knowledge. RESULTS: We evaluate SKEoG on two public datasets, that is, Chemical-Disease Relation (CDR) dataset and Chemical Reactions dataset (CHR) dataset, by comparing it with other state-of-the-art methods. SKEoG achieves the highest F1-score of 70.7 on the CDR dataset and F1-score of 91.4 on the CHR dataset. CONCLUSION: The proposed SKEoG method achieves new state-of-the-art performance. Both document structure and external knowledge can bring performance improvement in the EoG framework. Selecting proper methods for knowledge node representation is also very important.

CapsTM: capsule network for Chinese medical text matching.

BACKGROUND: Text Matching (TM) is a fundamental task of natural language processing widely used in many application systems such as information retrieval, automatic question answering, machine translation, dialogue system, reading comprehension, etc. In recent years, a large number of deep learning neural networks have been applied to TM, and have refreshed benchmarks of TM repeatedly. Among the deep learning neural networks, convolutional neural network (CNN) is one of the most popular networks, which suffers from difficulties in dealing with small samples and keeping relative structures of features. In this paper, we propose a novel deep learning architecture based on capsule network for TM, called CapsTM, where capsule network is a new type of neural network architecture proposed to address some of the short comings of CNN and shows great potential in many tasks. METHODS: CapsTM is a five-layer neural network, including an input layer, a representation layer, an aggregation layer, a capsule layer and a prediction layer. In CapsTM, two pieces of text are first individually converted into sequences of embeddings and are further transformed by a highway network in the input layer. Then, Bidirectional Long Short-Term Memory (BiLSTM) is used to represent each piece of text and attention-based interaction matrix is used to represent interactive information of the two pieces of text in the representation layer. Subsequently, the two kinds of representations are fused together by BiLSTM in the aggregation layer, and are further represented with capsules (vectors) in the capsule layer. Finally, the prediction layer is a connected network used for classification. CapsTM is an extension of ESIM by adding a capsule layer before the prediction layer. RESULTS: We construct a corpus of Chinese medical question matching, which contains 36,360 question pairs. This corpus is randomly split into three parts: a training set of 32,360 question pairs, a development set of 2000 question pairs and a test set of 2000 question pairs. On this corpus, we conduct a series of experiments to evaluate the proposed CapsTM and compare it with other state-of-the-art methods. CapsTM achieves the highest F-score of 0.8666. CONCLUSION: The experimental results demonstrate that CapsTM is effective for Chinese medical question matching and outperforms other state-of-the-art methods for comparison.