Search

Background and Objectives: This study aimed to confirm the effectiveness of the virtual passport for dementia care., Research Design and Methods: The virtual passport is an application (app) tool connecting healthcare professionals, dementia care sites, and people living with dementia (PLWD) and their family caregivers. This tool assists case managers in hospitals by providing individualized care plans and health education to PLWD and their caregivers. The dementia quality indicator achievement rates, care needs investigation and fulfillment, severity of behavioral and psychological symptoms of dementia (BPSD), and changes in caregiver burden and depression are measured at the initial interview and 6 and 12 months after the intervention., Results: We enrolled 57 and 54 patients and their caregivers in the virtual passport and routine care groups, respectively. Compared to the control group, six quality indicators in the passport group showed significantly higher achievement at 6 months after using the virtual passport. Case managers addressed more care needs at 6 months (1.37 vs 0, p < 0.001) and 12 months (1.32 vs 0, p < 0.001). Improvement in severity of neuropsychiatric symptoms (neuropsychiatric inventory (NPI) irritability/lability difference: -0.58 vs 0.22, p = 0.044; NPI agitation/aggression difference =-0.78 vs 0.00, p = 0.042) were also observed. No obvious influence was found in caregiver burden and depression after using the virtual passport., Discussion and Implications: The virtual passport is an effective information technology tool in improving the quality of dementia care, assisting case management in identifying more care needs, and reducing the severity of BPSD., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Hong, Wang, Yang, Tung, Dai, Hsu, Huang and Jhang.)

Background: International Classification of Diseases codes are widely used to describe diagnosis information, but manual coding relies heavily on human interpretation, which can be expensive, time consuming, and prone to errors. With the transition from the International Classification of Diseases, Ninth Revision, to the International Classification of Diseases, Tenth Revision (ICD-10), the coding process has become more complex, highlighting the need for automated approaches to enhance coding efficiency and accuracy. Inaccurate coding can result in substantial financial losses for hospitals, and a precise assessment of outcomes generated by a natural language processing (NLP)-driven autocoding system thus assumes a critical role in safeguarding the accuracy of the Taiwan diagnosis related groups (Tw-DRGs)., Objective: This study aims to evaluate the feasibility of applying an International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM), autocoding system that can automatically determine diagnoses and codes based on free-text discharge summaries to facilitate the assessment of Tw-DRGs, specifically principal diagnosis and major diagnostic categories (MDCs)., Methods: By using the patient discharge summaries from Kaohsiung Medical University Chung-Ho Memorial Hospital (KMUCHH) from April 2019 to December 2020 as a reference data set we developed artificial intelligence (AI)-assisted ICD-10-CM coding systems based on deep learning models. We constructed a web-based user interface for the AI-assisted coding system and deployed the system to the workflow of the certified coding specialists (CCSs) of KMUCHH. The data used for the assessment of Tw-DRGs were manually curated by a CCS with the principal diagnosis and MDC was determined from discharge summaries collected at KMUCHH from February 2023 to April 2023., Results: Both the reference data set and real hospital data were used to assess performance in determining ICD-10-CM coding, principal diagnosis, and MDC for Tw-DRGs. Among all methods, the GPT-2 (OpenAI)-based model achieved the highest F 1 -score, 0.667 (F 1 -score 0.851 for the top 50 codes), on the KMUCHH test set and a slightly lower F 1 -score, 0.621, in real hospital data. Cohen κ evaluation for the agreement of MDC between the models and the CCS revealed that the overall average κ value for GPT-2 (κ=0.714) was approximately 12.2 percentage points higher than that of the hierarchy attention network (κ=0.592). GPT-2 demonstrated superior agreement with the CCS across 6 categories of MDC, with an average κ value of approximately 0.869 (SD 0.033), underscoring the effectiveness of the developed AI-assisted coding system in supporting the work of CCSs., Conclusions: An NLP-driven AI-assisted coding system can assist CCSs in ICD-10-CM coding by offering coding references via a user interface, demonstrating the potential to reduce the manual workload and expedite Tw-DRG assessment. Consistency in performance affirmed the effectiveness of the system in supporting CCSs in ICD-10-CM coding and the judgment of Tw-DRGs., (©Hong-Jie Dai, Chen-Kai Wang, Chien-Chang Chen, Chong-Sin Liou, An-Tai Lu, Chia-Hsin Lai, Bo-Tsz Shain, Cheng-Rong Ke, William Yu Chung Wang, Tatheer Hussain Mir, Mutiara Simanjuntak, Hao-Yun Kao, Ming-Ju Tsai, Vincent S Tseng. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 20.09.2024.)

Type 2 diabetes mellitus (T2DM) is a risk factor for patients with impaired renal function. The onset of T2DM-induced diabetic kidney disease (DKD) is frequently sub-clinical, potentially culminating in end-stage renal disease. In the current study the factors influencing DKD in elderly patients diagnosed with T2DM were determined. A retrospective cohort study was conducted involving patients ≥60 years of age with T2DM from June 2019 to December 2022. The Cockcroft-Gault formula was used to estimate the glomerular filtration rate. The clinical information and biochemical indicators of patients with an estimated glomerular filtration rate (eGFR) < 90 mL/min/1.73m2 were collected. Patients were grouped based on a 3-year eGFR decline < 15% and ≥ 15%. The differences between the two groups were compared and the factors influencing the 3-year eGFR decline ≥ 15% were analyzed. A total of 242 patients were included, including 154 in the group with a 3-year eGFR decline < 15% and 88 in the group with a three-year eGFR decline ≥ 15%. Univariate logistic regression analysis showed that smoking cigarettes, and triglycerides (TG) and high-density lipoprotein levels were related to a 3-year eGFR decline ≥ 15% (P = .039, P < .001, and P = .011, respectively). Multivariate logistic regression analysis showed that the TG level was independently related to a 3-year eGFR decline ≥ 15% (P = .004; OR = 2.316). There was a significant linear relationship between the eGFR decline and TG level (P = .002). Patients with a TG concentration > 1.7 mmol/L had a more apparent decrease in the eGFR (P < .05). For elderly patients with T2DM and an eGFR < 90 mL/min/1.73m2, the TG level may be an important risk factor for deteriorating renal function that warrants actively intervention., Competing Interests: The authors have no conflicts of interest to disclose., (Copyright © 2024 the Author(s). Published by Wolters Kluwer Health, Inc.)

Data curation for a hospital-based cancer registry heavily relies on the labor-intensive manual abstraction process by cancer registrars to identify cancer-related information from free-text electronic health records. To streamline this process, a natural language processing system incorporating a hybrid of deep learning-based and rule-based approaches for identifying lung cancer registry-related concepts, along with a symbolic expert system that generates registry coding based on weighted rules, was developed. The system is integrated with the hospital information system at a medical center to provide cancer registrars with a patient journey visualization platform. The embedded system offers a comprehensive view of patient reports annotated with significant registry concepts to facilitate the manual coding process and elevate overall quality. Extensive evaluations, including comparisons with state-of-the-art methods, were conducted using a lung cancer dataset comprising 1428 patients from the medical center. The experimental results illustrate the effectiveness of the developed system, consistently achieving F1-scores of 0.85 and 1.00 across 30 coding items. Registrar feedback highlights the system's reliability as a tool for assisting and auditing the abstraction. By presenting key registry items along the timeline of a patient's reports with accurate code predictions, the system improves the quality of registrar outcomes and reduces the labor resources and time required for data abstraction. Our study highlights advancements in cancer registry coding practices, demonstrating that the proposed hybrid weighted neural-symbolic cancer registry system is reliable and efficient for assisting cancer registrars in the coding workflow and contributing to clinical outcomes., Competing Interests: The authors declare no conflict of interest., (© 2024 The Authors.)

Introduction: The clinical presentations of dry eye disease (DED) and depression (DEP) often comanifest. However, the robustness and the mechanisms underlying this association were undetermined., Objectives: To this end, we set up a three-segment study that employed multimodality results (meta-analysis, genome-wide association study [GWAS] and Mendelian randomization [MR]) to elucidate the association, common pathways and causality between DED and DEP., Methods: A meta-analysis comprising 26 case-control studies was first conducted to confirm the DED-DEP association. Next, we performed a linkage disequilibrium (LD)-adjusted GWAS and targeted phenotype association study (PheWAS) in East Asian TW Biobank (TWB) and European UK Biobank (UKB) populations. Single-nucleotide polymorphisms (SNPs) were further screened for molecular interactions and common pathways at the functional gene level. To further elucidate the activated pathways in DED and DEP, a systemic transcriptome review was conducted on RNA sequencing samples from the Gene Expression Omnibus. Finally, 48 MR experiments were implemented to examine the bidirectional causation between DED and DEP., Results: Our meta-analysis showed that DED patients are associated with an increased DEP prevalence (OR = 1.83), while DEP patients have a concurrent higher risk of DED (OR = 2.34). Notably, cross-disease GWAS analysis revealed that similar genetic architecture (r G  = 0.19) and pleiotropic functional genes contributed to phenotypes in both diseases. Through protein-protein interaction and ontology convergence, we summarized the pleiotropic functional genes under the ontology of immune activation, which was further validated by a transcriptome systemic review. Importantly, the inverse variance-weighted (IVW)-MR experiments in both TWB and UKB populations (p value <0.001) supported the bidirectional exposure-outcome causation for DED-to-DEP and DEP-to-DED. Despite stringent LD-corrected instrumental variable re-selection, the bidirectional causation between DED and DEP remained., Conclusion: With the multi-modal evidence combined, we consolidated the association and causation between DED and DEP., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023. Production and hosting by Elsevier B.V.)

Background: The widespread use of electronic health records in the clinical and biomedical fields makes the removal of protected health information (PHI) essential to maintain privacy. However, a significant portion of information is recorded in unstructured textual forms, posing a challenge for deidentification. In multilingual countries, medical records could be written in a mixture of more than one language, referred to as code mixing. Most current clinical natural language processing techniques are designed for monolingual text, and there is a need to address the deidentification of code-mixed text., Objective: The aim of this study was to investigate the effectiveness and underlying mechanism of fine-tuned pretrained language models (PLMs) in identifying PHI in the code-mixed context. Additionally, we aimed to evaluate the potential of prompting large language models (LLMs) for recognizing PHI in a zero-shot manner., Methods: We compiled the first clinical code-mixed deidentification data set consisting of text written in Chinese and English. We explored the effectiveness of fine-tuned PLMs for recognizing PHI in code-mixed content, with a focus on whether PLMs exploit naming regularity and mention coverage to achieve superior performance, by probing the developed models' outputs to examine their decision-making process. Furthermore, we investigated the potential of prompt-based in-context learning of LLMs for recognizing PHI in code-mixed text., Results: The developed methods were evaluated on a code-mixed deidentification corpus of 1700 discharge summaries. We observed that different PHI types had preferences in their occurrences within the different types of language-mixed sentences, and PLMs could effectively recognize PHI by exploiting the learned name regularity. However, the models may exhibit suboptimal results when regularity is weak or mentions contain unknown words that the representations cannot generate well. We also found that the availability of code-mixed training instances is essential for the model's performance. Furthermore, the LLM-based deidentification method was a feasible and appealing approach that can be controlled and enhanced through natural language prompts., Conclusions: The study contributes to understanding the underlying mechanism of PLMs in addressing the deidentification process in the code-mixed context and highlights the significance of incorporating code-mixed training instances into the model training phase. To support the advancement of research, we created a manipulated subset of the resynthesized data set available for research purposes. Based on the compiled data set, we found that the LLM-based deidentification method is a feasible approach, but carefully crafted prompts are essential to avoid unwanted output. However, the use of such methods in the hospital setting requires careful consideration of data security and privacy concerns. Further research could explore the augmentation of PLMs and LLMs with external knowledge to improve their strength in recognizing rare PHI., (©You-Qian Lee, Ching-Tai Chen, Chien-Chang Chen, Chung-Hong Lee, Peitsz Chen, Chi-Shin Wu, Hong-Jie Dai. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 25.01.2024.)

Background: Electronic health records (EHRs) in unstructured formats are valuable sources of information for research in both the clinical and biomedical domains. However, before such records can be used for research purposes, sensitive health information (SHI) must be removed in several cases to protect patient privacy. Rule-based and machine learning-based methods have been shown to be effective in deidentification. However, very few studies investigated the combination of transformer-based language models and rules., Objective: The objective of this study is to develop a hybrid deidentification pipeline for Australian EHR text notes using rules and transformers. The study also aims to investigate the impact of pretrained word embedding and transformer-based language models., Methods: In this study, we present a hybrid deidentification pipeline called OpenDeID, which is developed using an Australian multicenter EHR-based corpus called OpenDeID Corpus. The OpenDeID corpus consists of 2100 pathology reports with 38,414 SHI entities from 1833 patients. The OpenDeID pipeline incorporates a hybrid approach of associative rules, supervised deep learning, and pretrained language models., Results: The OpenDeID achieved a best F 1 -score of 0.9659 by fine-tuning the Discharge Summary BioBERT model and incorporating various preprocessing and postprocessing rules. The OpenDeID pipeline has been deployed at a large tertiary teaching hospital and has processed over 8000 unstructured EHR text notes in real time., Conclusions: The OpenDeID pipeline is a hybrid deidentification pipeline to deidentify SHI entities in unstructured EHR text notes. The pipeline has been evaluated on a large multicenter corpus. External validation will be undertaken as part of our future work to evaluate the effectiveness of the OpenDeID pipeline., (©Jiaxing Liu, Shalini Gupta, Aipeng Chen, Chen-Kai Wang, Pratik Mishra, Hong-Jie Dai, Zoie Shui-Yee Wong, Jitendra Jonnagaddala. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 06.12.2023.)

The contamination of food by pathogens is a serious problem in global food safety, and current methods of detection are costly, time-consuming, and cumbersome. Therefore, it is necessary to develop rapid, portable, and sensitive assays for foodborne pathogens. In addition, assays for foodborne pathogens must be resistant to interference resulting from the complex food matrix to prevent false positives and negatives. In this study, hemin and reduced graphene oxide-MoS 2 sheets (GMS) were used to design a near-infrared (NIR)-responsive photoelectrochemical (PEC) aptasensor with target-induced photocurrent polarity switching based on a hairpin aptamer (Hp) with a G-quadruplex motif. A ready-to-use analytical device was developed by immobilizing GMS on the surface of a commercial screen-printed electrode, followed by the attachment of the aptamer. In the presence of Escherichia coli O157:H7, the binding sites of Hp with the G-quadruplex motif were opened and exposed to hemin, leading to the formation of a G-quadruplex/hemin DNAzyme. Crucially, after binding to hemin, the charge transfer pathway of GMS changes, resulting in a switch of the photocurrent polarity. Further, G-quadruplex/hemin DNAzyme enhanced the cathodic photocurrent, and the proposed sensor exhibited a wide linear range ((25.0-1.0) × 10 7 CFU/mL), a low limit of detection (2.0 CFU/mL), and good anti-interference performance. These findings expand the applications of NIR-responsive PEC materials and provide versatile PEC methods for detecting biological analytes, especially for food safety testing.

This study presents the outcomes of the shared task competition BioCreative VII (Task 3) focusing on the extraction of medication names from a Twitter user's publicly available tweets (the user's 'timeline'). In general, detecting health-related tweets is notoriously challenging for natural language processing tools. The main challenge, aside from the informality of the language used, is that people tweet about any and all topics, and most of their tweets are not related to health. Thus, finding those tweets in a user's timeline that mention specific health-related concepts such as medications requires addressing extreme imbalance. Task 3 called for detecting tweets in a user's timeline that mentions a medication name and, for each detected mention, extracting its span. The organizers made available a corpus consisting of 182 049 tweets publicly posted by 212 Twitter users with all medication mentions manually annotated. The corpus exhibits the natural distribution of positive tweets, with only 442 tweets (0.2%) mentioning a medication. This task was an opportunity for participants to evaluate methods that are robust to class imbalance beyond the simple lexical match. A total of 65 teams registered, and 16 teams submitted a system run. This study summarizes the corpus created by the organizers and the approaches taken by the participating teams for this challenge. The corpus is freely available at https://biocreative.bioinformatics.udel.edu/tasks/biocreative-vii/track-3/. The methods and the results of the competing systems are analyzed with a focus on the approaches taken for learning from class-imbalanced data., (© The Author(s) 2023. Published by Oxford University Press.)

Background: Most previous studies make psychiatric diagnoses based on diagnostic terms. In this study we sought to augment Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) diagnostic criteria with deep neural network models to make psychiatric diagnoses based on psychiatric notes., Methods: We augmented DSM-5 diagnostic criteria with self-attention-based bidirectional long short-term memory (BiLSTM) models to identify schizophrenia, bipolar, and unipolar depressive disorders. Given that the diagnostic criteria for psychiatric diagnosis include a certain symptom profile and functional impairment, we first extracted psychiatric symptoms and functional features with two approaches, including a lexicon-based approach and a dependency parsing approach. Then, we incorporated free-text discharge notes and extracted features for psychiatric diagnoses with the proposed models., Results: The micro-averaged F1 scores of the two automatic annotation approaches were greater than 0.8. BiLSTM models with self-attention outperformed the rule-based models with DSM-5 criteria in the prediction of schizophrenia and bipolar disorder, while the latter outperformed the former in predicting unipolar depressive disorder. Approaches for augmenting DSM-5 criteria with a self-attention-based BiLSTM outperformed both pure rule-based and pure deep neural network models. In terms of classification of psychiatric diagnoses, we observed that the performance for schizophrenia and bipolar disorder was acceptable., Conclusion: This DSM-5-augmented deep neural network models showed good performance in identifying psychiatric diagnoses from psychiatric notes. We conclude that it is possible to establish a model that consults clinical notes to make psychiatric diagnoses comparably to physicians. Further research will be extended to outpatient notes and other psychiatric disorders., Competing Interests: Declaration of competing interest The authors have no relevant financial or non-financial interests, or competing interests to disclose., (Copyright © 2023. Published by Elsevier B.V.)

Ventilator-induced lung injury (VILI) is a lung injury induced or aggravated by mechanical ventilation. Transforming growth factor (TGF)-β1 is a cytokine that mediates immune function, enabling inflammatory attenuation and tissue repair. Here, we hypothesized that it plays an important role in the attenuation of VILI and inflammation. Ventilation with high tidal volume was performed on C57BL/6 mice to establish a VILI model. After 4 h of ventilation, mice were sacrificed (end of ventilation [EOV]) or extubated for resuscitation at 4 h (post-ventilation 4 h [PV4h]), 8 h (PV8h) and 24 h post-ventilation (PV1d). Recombinant mouse TGF-β1 (rTGF-β1) and the neutralization antibody of TGF-β1 (nTAb) were used in vivo to examine the effect of TGF-β1 on immune function and inflammatory attenuation in VILI mice. Lung injury was exacerbated at the same trend as the interleukin (IL)-1β level, peaking at PV1d, whereas IL-6 and tumor necrosis factor (TNF)-α levels gradually reduced. Most active phagosomes, swollen round mitochondria, and cavitating lamellar bodies were observed at PV4h. The CD4+ T cells were significantly increased from PV4h to PV1d, and the CD8a + T cells were higher in the PV4h and PV1d groups; furthermore, the mice in the PV8h group showed highest proportion of CD4+CD8a+ T cells and CD4+/CD8a+ ratio. CD19 + and CD5 + CD19 + B cells in VILI mice began to increase at PV1d. The pulmonary expression of latent and monomer TGF-β1 increased at PV4h and PV8h. Treatment of rTGF-β1 only induced high expression of latent and monomer TGF-β1 at EOV to decrease pulmonary levels of IL-1β, IL-6, and TNF-α; however, lung injury attenuated from EOV to PV1d. TGF-β1 induced the delayed elevation of CD4+/CD8a+ T cells ratio and activation of pulmonary CD4+CD8a+ double-positive T cells under certain conditions. Elastic fibers and celluloses, although relatively less proteoglycan, were observed with the overexpression of TGF-β1 at PV4h and PV8h. In conclusion, TGF-β1 attenuates the inflammatory response and lung injury of VILI via immune function regulation., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2022 Elsevier B.V. All rights reserved.)

Cerebral visual impairments (CVIs) is an umbrella term that categorizes miscellaneous visual defects with parallel genetic brain disorders. While the manifestations of CVIs are diverse and ambiguous, molecular diagnostics stand out as a powerful approach for understanding pathomechanisms in CVIs. Nevertheless, the characterization of CVI disease cohorts has been fragmented and lacks integration. By revisiting the genome-wide and phenome-wide association studies (GWAS and PheWAS), we clustered a handful of renowned CVIs into five ontology groups, namely ciliopathies (Joubert syndrome, Bardet-Biedl syndrome, Alstrom syndrome), demyelination diseases (multiple sclerosis, Alexander disease, Pelizaeus-Merzbacher disease), transcriptional deregulation diseases (Mowat-Wilson disease, Pitt-Hopkins disease, Rett syndrome, Cockayne syndrome, X-linked alpha-thalassaemia mental retardation), compromised peroxisome disorders (Zellweger spectrum disorder, Refsum disease), and channelopathies (neuromyelitis optica spectrum disorder), and reviewed several mutation hotspots currently found to be associated with the CVIs. Moreover, we discussed the common manifestations in the brain and the eye, and collated animal study findings to discuss plausible gene editing strategies for future CVI correction.

Background: Metabolic dysfunction-associated fatty liver disease (MAFLD) is a newly proposed definition of fatty liver disease (FLD) independent of excessive alcohol consumption (EAC) and hepatitis viral infection. Evidence on the mortality risk in different types of FLD [nonalcoholic FLD (NAFLD), alcoholic FLD (AFLD), and MAFLD] is sparse, hindering the identification of high-risk populations for preferential clinical surveillance., Methods: A total of 11,000 participants in the Third National Health and Nutrition Examination Survey were enrolled. Participants were categorized into three groups [FLD( - ), MAFLD( - ), and MAFLD( +)] according to FLD and MAFLD criteria, and further categorized into six groups by EAC. Multivariate Cox proportional hazard model was used to estimate the risk of all-cause, cardiovascular-related, and cancer-related mortality., Results: During a median follow-up of 23.2 years, a total of 3240 deaths were identified. Compared with FLD( - )/EAC( - ) participants, MAFLD( +) individuals had higher all-cause mortality risk [hazard ratio (HR) = 1.28, 95% confidence interval (CI) = 1.18-1.39] regardless of EAC status [MAFLD( +)/NAFLD: HR = 1.22, 95%CI = 1.11-1.34; MAFLD( +)/AFLD: HR = 1.83, 95%CI = 1.46-2.28], while not for MAFLD( - ) individuals. Furthermore, diabetes-driven-MAFLD had higher mortality risk (HR = 2.00, 95%CI = 1.77-2.27) followed by metabolic dysregulation-driven-MAFLD (HR = 1.30, 95%CI = 1.06-1.60) and overweight/obesity-driven-MAFLD (HR = 1.11, 95%CI = 1.00-1.22). Additionally, MAFLD( - ) participants with elevated fibrosis score were also associated with statistically significantly higher mortality risk (HR = 3.23, 95%CI = 1.63-6.40)., Conclusions: Utilizing a representative sample of the US population, we proved the validity of MAFLD subtype and fibrosis score, rather than the traditional definition (NAFLD and AFLD), in the risk stratification of FLD patients. These findings may be applied to guide the determination of surveillance options for FLD patients., (© 2022. Asian Pacific Association for the Study of the Liver.)

Electronic health records (EHRs) at medical institutions provide valuable sources for research in both clinical and biomedical domains. However, before such records can be used for research purposes, protected health information (PHI) mentioned in the unstructured text must be removed. In Taiwan's EHR systems the unstructured EHR texts are usually represented in the mixing of English and Chinese languages, which brings challenges for de-identification. This paper presented the first study, to the best of our knowledge, of the construction of a code-mixed EHR de-identification corpus and the evaluation of different mature entity recognition methods applied for the code-mixed PHI recognition task.

The heightened cardiovascular disease (CVD) risk observed among omnivores is thought to be linked, in part, to gut microbiota-dependent generation of trimethylamine-N-oxide (TMAO) from L-carnitine, a nutrient abundant in red meat. Gut microbial transformation of L-carnitine into trimethylamine (TMA), the precursor of TMAO, occurs via the intermediate γ-butyrobetaine (γBB). However, the interrelationship of γBB, red meat ingestion and CVD risks, as well as the gut microbial genes responsible for the transformation of γBB to TMA, are unclear. In the present study, we show that plasma γBB levels in individuals from a clinical cohort (n = 2,918) are strongly associated with incident CVD event risks. Culture of human faecal samples and microbial transplantation studies in gnotobiotic mice with defined synthetic communities showed that the introduction of Emergencia timonensis, a human gut microbe that can metabolize γBB into TMA, is sufficient to complete the carnitine → γBB → TMA transformation, elevate TMAO levels and enhance thrombosis potential in recipients after arterial injury. RNA-sequencing analyses of E. timonensis identified a six-gene cluster, herein named the γBB utilization (gbu) gene cluster, which is upregulated in response to γBB. Combinatorial cloning and functional studies identified four genes (gbuA, gbuB, gbuC and gbuE) that are necessary and sufficient to recapitulate the conversion of γBB to TMA when coexpressed in Escherichia coli. Finally, reanalysis of samples (n = 113) from a clinical, randomized diet, intervention study showed that the abundance of faecal gbuA correlates with plasma TMAO and a red meat-rich diet. Our findings reveal a microbial gene cluster that is critical to dietary carnitine → γBB → TMA → TMAO transformation in hosts and contributes to CVD risk., (© 2021. The Author(s), under exclusive licence to Springer Nature Limited.)

The introduction of pre-trained language models in natural language processing (NLP) based on deep learning and the availability of electronic health records (EHRs) presents a great opportunity to transfer the "knowledge" learned from data in the general domain to enable the analysis of unstructured textual data in clinical domains. This study explored the feasibility of applying NLP to a small EHR dataset to investigate the power of transfer learning to facilitate the process of patient screening in psychiatry. A total of 500 patients were randomly selected from a medical center database. Three annotators with clinical experience reviewed the notes to make diagnoses for major/minor depression, bipolar disorder, schizophrenia, and dementia to form a small and highly imbalanced corpus. Several state-of-the-art NLP methods based on deep learning along with pre-trained models based on shallow or deep transfer learning were adapted to develop models to classify the aforementioned diseases. We hypothesized that the models that rely on transferred knowledge would be expected to outperform the models learned from scratch. The experimental results demonstrated that the models with the pre-trained techniques outperformed the models without transferred knowledge by micro-avg. and macro-avg. F-scores of 0.11 and 0.28, respectively. Our results also suggested that the use of the feature dependency strategy to build multi-labeling models instead of problem transformation is superior considering its higher performance and simplicity in the training process., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2021 Dai, Su, Lee, Zhang, Wang, Kuo and Wu.)

Background: Identifying and extracting family history information (FHI) from clinical reports are significant for recognizing disease susceptibility. However, FHI is usually described in a narrative manner within patients' electronic health records, which requires the application of natural language processing technologies to automatically extract such information to provide more comprehensive patient-centered information to physicians., Objective: This study aimed to overcome the 2 main challenges observed in previous research focusing on FHI extraction. One is the requirement to develop postprocessing rules to infer the member and side information of family mentions. The other is to efficiently utilize intrasentence and intersentence information to assist FHI extraction., Methods: We formulated the task as a sequential labeling problem and propose an enhanced relation-side scheme that encodes the required family member properties to not only eliminate the need for postprocessing rules but also relieve the insufficient training instance issues. Moreover, an attention-based neural network structure was proposed to exploit cross-sentence information to identify FHI and its attributes requiring cross-sentence inference., Results: The dataset released by the 2019 n2c2/OHNLP family history extraction task was used to evaluate the performance of the proposed methods. We started by comparing the performance of the traditional neural sequence models with the ordinary scheme and enhanced scheme. Next, we studied the effectiveness of the proposed attention-enhanced neural networks by comparing their performance with that of the traditional networks. It was observed that, with the enhanced scheme, the recall of the neural network can be improved, leading to an increase in the F score of 0.024. The proposed neural attention mechanism enhanced both the recall and precision and resulted in an improved F score of 0.807, which was ranked fourth in the shared task., Conclusions: We presented an attention-based neural network along with an enhanced tag scheme that enables the neural network model to learn and interpret the implicit relationship and side information of the recognized family members across sentences without relying on heuristic rules., (©Hong-Jie Dai, You-Qian Lee, Chandini Nekkantti, Jitendra Jonnagaddala. Originally published in JMIR Medical Informatics (http://medinform.jmir.org), 01.12.2020.)

Background: In animal models of ventilation-induced lung injury, mitophagy triggers mitochondria damage and the release of mitochondrial (mt) DNA, which activates inflammation. However, the mechanism of this process is unclear., Methods: A model of cyclic stretching (CS)-induced lung epithelial cell injury was established. The genetic intervention of phosphatase and tensin homolog-induced kinase 1 (PINK1) expression via lentivirus transfection was used to identify the relationship between PINK1-mediated mitophagy and mtDNA release in stretching-induced inflammatory response and injury. Pharmacological inhabitation of Toll-like receptor 9 (TLR9) and myeloid differentiation factor 88 (MyD88) expression was performed via their related inhibitors, while pre-treatment of exogenous mtDNA was used to verify the role of mtDNA in stretching-induced inflammatory response and injury., Results: Using a cell culture model of CS, we found that knocking down PINK1 in lung epithelial cells reduced mitophagy activation and mtDNA release, leading to milder inflammatory response and injury; conversely, up-regulating PINK1 exacerbated stretching-induced inflammation and injury, and similar effects were observed by upregulating TLR9 to induce expression of MyD88 and nuclear factor-κB (NF-κB)/p65. Down-regulating MyD88 protected lung epithelial cells from stretching injury and decreased NF-κB/p65 expression., Conclusion: These findings suggest that PINK1-dependent mitophagy and associated TLR9 activation is indeed a major factor in stretch-induced cell injury via a mechanism in which released mtDNA activates TLR9 and thereby the MyD88/NF-κB pathway. Inhibiting this process may be a therapeutic approach to prevent inflammation and cell injury in patients on mechanical ventilation., (Copyright © 2020 Jing, Hu, Lin, He, Zhang, Ge, Dai, Du, Lin and Pan.)

Background: The preoperative presence of diabetes mellitus (DM) has been recently demonstrated to be a risk factor for adverse events after thoracic surgery. However, the specific effects of presence of DM preoperatively on thoracic surgery is not known. This study aimed to investigate the association between preoperative DM and clinical outcomes and the short-term survival rates after thoracic surgery., Methods: In this retrospective, observational, and matched-pair analysis study, patients receiving thoracic surgery from a tertiary university hospital in 2 consecutive years were grouped as either type 2 DM (T2DM) or controlled within the first 24 hours after surgery. Multivariate Cox regression was conducted to investigate the impact of T2DM within the first 24 hours of admission on in-intensive care unit (ICU) and hospital survival., Results: Among the included thoracic patients, 41 (8.4%) had T2DM and 450 (91.6%) did not have T2DM. In the single-factor analyses, T2DM patients were shown to have a higher preoperative white blood cells (WBCs) count, increased release of immunoglobulin A, complement C3 and C4, impaired kidney function with high level of urea, and low expression of alanine aminotransferase (ALT) and monoamine oxidase (MAO). In multivariate analyses, the preoperative urea level was associated with a low-grade risk of dying for the ICU survival time. In contrast, preoperative complement C3 level favored a positive contribution in-ICU survival. Besides the complement C3 level, immunoglobulin A level remained a positive contribution in regression models of hospital survival., Conclusions: Pre-admission T2DM was not associated with an increased in-ICU and hospital mortality among patients with thoracic surgery. Furthermore, they were accompanied by impaired kidney function, activated inflammation and liver function.

The ataxia-telangiectasia mutated (ATM) protein is regarded as the linchpin of cellular defenses to stress. Deletion of ATM results in strong oxidative stress and degenerative diseases in the nervous system. However, the role of ATM in neuronal ischemic preconditioning and lethal ischemic injury is still largely unknown. In this study, mice cortical neurons preconditioned with sublethal exposure to oxygen glucose deprivation (OGD) exhibited ATM/glucose-6-phosphate dehydrogenase pathway activation. Additionally, pharmacological inhibition of ATM prior to the preconditioning reversed neuroprotection provided by preconditioning in vitro and in vivo. Meanwhile, we found that ATM/P53 pro-apoptosis pathway was driven by lethal OGD injury, and pharmacological inhibition of ATM during fatal oxygen-glucose deprivation/reperfusion injury promoted neuronal survival. More importantly, inhibition of ATM activity after cerebral ischemia protected against cerebral ischemic-reperfusion damage in mice. In conclusion, our data show the dual role of ATM in neuronal ischemic preconditioning and lethal ischemic injury, involving in the protection of ischemic preconditioning, but promoting neuronal death in lethal ischemic injury. Thus, the present study provides new opportunity for the treatment of ischemic stroke.

Identifying patients eligible for clinical trials using electronic health records (EHRs) is a challenging task usually requiring a comprehensive analysis of information stored in multiple EHRs of a patient. The goal of this study is to investigate different methods and their effectiveness in identifying patients that meet specific eligibility selection criteria based on patients' longitudinal records. An unstructured dataset released by the n2c2 cohort selection for clinical trials track was used, each of which included 2-5 records manually annotated to thirteen pre-defined selection criteria. Unlike the other studies, we formulated the problem as a multiple instance learning (MIL) task and compared the performance with that of the rule-based and the single instance-based classifiers. Our official best run achieved an average micro-F score of 0.8765 which was ranked as one of the top ten results in the track. Further experiments demonstrated that the performance of the MIL-based classifiers consistently yield better performance than their single-instance counterparts in the criteria that require the overall comprehension of the information distributed among all of the patient's EHRs. Rule-based and single instance learning approaches exhibited better performance in criteria that don't require a consideration of several factors across records. This study demonstrated that cohort selection using longitudinal patient records can be formulated as a MIL problem. Our results exhibit that the MIL-based classifiers supplement the rule-based methods and provide better results in comparison to the single instance learning approaches., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2020 Elsevier Inc. All rights reserved.)

Objective: Liver fibrosis is an important predictor of mortality in nonalcoholic fatty liver disease (NAFLD). Peripheral artery disease (PAD) and liver fibrosis share many common metabolic dysfunctions. We aimed to explore the association between PAD and risk of fibrosis deterioration in NAFLD patients., Methods: The study recruited 1,610 NAFLD patients aged ≥ 40 years from a well-defined community at baseline in 2010 and followed up between August 2014 and May 2015. Fibrosis deterioration was defined as the NAFLD fibrosis score (NFS) status increased to a higher category at the follow-up visit. PAD was defined as an ankle-brachial index of < 0.90 or > 1.40., Results: During an average of 4.3 years' follow-up, 618 patients progressed to a higher NFS category. PAD was associated with 92% increased risk of fibrosis deterioration [multivariable-adjusted odds ratio ( OR ): 1.92, 95% confidence interval ( CI ): 1.24, 2.98]. When stratified by baseline NFS status, the OR for progression from low to intermediate or high NFS was 1.74 (95% CI : 1.02, 3.00), and progression from intermediate to high NFS was 2.24 (95% CI : 1.05, 4.80). There was a significant interaction between PAD and insulin resistance (IR) on fibrosis deterioration ( P for interaction = 0.03). As compared with non-PAD and non-IR, the coexistence of PAD and IR was associated with a 3.85-fold (95% CI : 2.06, 7.18) increased risk of fibrosis deterioration., Conclusion: PAD is associated with an increased risk of fibrosis deterioration in NAFLD patients, especially in those with IR. The coexistence of PAD and IR may impose an interactive effect on the risk of fibrosis deterioration., (Copyright © 2020 The Editorial Board of Biomedical and Environmental Sciences. Published by China CDC. All rights reserved.)

Purpose: Previous studies revealed that high serum uric acid (SUA) levels correlated with increased triglyceride to high-density lipoprotein cholesterol (TG/HDL-C) ratio. However, evidence is lacking regarding whether TG/HDL-C is an independent risk factor or just a simple marker of hyperuricemia. We aimed to investigate the relationship between TG/HDL-C and the risk of hyperuricemia in Chinese population., Methods: This retrospective cohort study involved 15,198 subjects (43.14 ± 13.13 years, 54.9% men) who were free of hyperuricemia at baseline. The association between TG/HDL-C and the risk of hyperuricemia, in the total sample and stratified by subgroups, was examined by multiple logistic regression analyses., Results: During 4 years follow-up, hyperuricemia occurred in 2365 (15.6%) participants. The cumulative incidence of hyperuricemia was significantly elevated with increasing TG/HDL-C quartiles (5.9, 10.8, 18.4, and 27.5%, respectively). After multivariate adjustment, the odds ratio for hyperuricemia in the highest compared with the lowest quartile of TG/HDL-C was 1.80 (95% confidence interval [CI] 1.49, 2.18), and each SD increment of TG/HDL-C ratio caused a 10% additional risk for hyperuricemia. Moreover, subgroup analyses showed that the association between TG/HDL-C and the risk of hyperuricemia was more pronounced in women and normal-weight adults. The results were consistent when analyses were restricted to participants without metabolic syndrome., Conclusions: TG/HDL-C ratio was positively related to the risk of hyperuricemia in Chinese population, particularly in women and normal-weight individuals. These findings suggested the potential of TG/HDL-C ratio to serve as an independent risk indicator in the prevention of hyperuricemia.

Background: Many studies have used Taiwan's National Health Insurance Research database (NHIRD) to conduct psychiatric research. However, the accuracy of the diagnostic codes for psychiatric disorders in NHIRD is not validated, and the symptom profiles are not available either. This study aimed to evaluate the accuracy of diagnostic codes and use text mining to extract symptom profile and functional impairment from electronic health records (EHRs) to overcome the above research limitations., Methods: A total of 500 discharge notes were randomly selected from a medical center's database. Three annotators reviewed the notes to establish gold standards. The accuracy of diagnostic codes for major psychiatric illness was evaluated. Text mining approaches were applied to extract depressive symptoms and function profiles and to identify patients with major depressive disorder., Results: The accuracy of the diagnostic code for major depressive disorder, schizophrenia, and dementia was acceptable but that of bipolar disorder and minor depression was less satisfactory. The performance of text mining approach to recognize depressive symptoms is satisfactory; however, the recall for functional impairment is lower resulting in lower F-scores of 0.774-0.753. Using the text mining approach to identify major depressive disorder, the recall was 0.85 but precision was only 0.69., Conclusions: The accuracy of the diagnostic code for major depressive disorder in discharge notes was generally acceptable. This finding supports the utilization of psychiatric diagnoses in claims databases. The application of text mining to EHRs might help in overcoming current limitations in research using claims databases., (Copyright © 2019. Published by Elsevier B.V.)

Objective: An adverse drug event (ADE) refers to an injury resulting from medical intervention related to a drug including harm caused by drugs or from the usage of drugs. Extracting ADEs from clinical records can help physicians associate adverse events to targeted drugs., Materials and Methods: We proposed a cascading architecture to recognize medical concepts including ADEs, drug names, and entities related to drugs. The architecture includes a preprocessing method and an ensemble of conditional random fields (CRFs) and neural network-based models to respectively address the challenges of surrogate string and overlapping annotation boundaries observed in the employed ADEs and medication extraction (ADME) corpus. The effectiveness of applying different pretrained and postprocessed word embeddings for the ADME task was also studied., Results: The empirical results showed that both CRFs and neural network-based models provide promising solution for the ADME task. The neural network-based models particularly outperformed CRFs in concept types involving narrative descriptions. Our best run achieved an overall micro F-score of 0.919 on the employed corpus. Our results also suggested that the Global Vectors for word representation embedding in general domain provides a very strong baseline, which can be further improved by applying the principal component analysis to generate more isotropic vectors., Conclusions: We have demonstrated that the proposed cascading architecture can handle the problem of overlapped annotations and further improve the overall recall and F-scores because the architecture enables the developed models to exploit more context information and forms an ensemble for creating a stronger recognizer., (© The Author(s) 2019. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Background: Family history information (FHI) described in unstructured electronic health records (EHRs) is a valuable information source for patient care and scientific researches. Since FHI is usually described in the format of free text, the entire process of FHI extraction consists of various steps including section segmentation, family member and clinical observation extraction, and relation discovery between the extracted members and their observations. The extraction step involves the recognition of FHI concepts along with their properties such as the family side attribute of the family member concept., Methods: This study focuses on the extraction step and formulates it as a sequence labeling problem. We employed a neural sequence labeling model along with different tag schemes to distinguish family members and their observations. Corresponding to different tag schemes, the identified entities were aggregated and processed by different algorithms to determine the required properties., Results: We studied the effectiveness of encoding required properties in the tag schemes by evaluating their performance on the dataset released by the BioCreative/OHNLP challenge 2018. It was observed that the proposed side scheme along with the developed features and neural network architecture can achieve an overall F1-score of 0.849 on the test set, which ranked second in the FHI entity recognition subtask., Conclusions: By comparing with the performance of conditional random fields models, the developed neural network-based models performed significantly better. However, our error analysis revealed two challenging issues of the current approach. One is that some properties required cross-sentence inferences. The other is that the current model is not able to distinguish between the narratives describing the family members of the patient and those specifying the relatives of the patient's family members.

Objective: To clarify the effect of TRAF2 in the biological behavior of gastric cancer and explore the mechanism. Methods: TRAF2 stably depleted AGS cell was established. Cell growth was monitored by x-CELLigence system. Cell proliferation was detected using cell viability assay. The apoptosis and cell cycle were detected by flow cytometry. The difference of migration and invasion abilities were measured by real-time xCELLigence system and Transwell. The expression and activity of NF-κB signaling pathway were measured by western blot and TransAM assay. The expression of TRAF2 in gastric cancer tissue and its clinical significance were detected by immunohistochemistry. Results: The cell index of AGS-siTRAF2 cells was significantly lower than that of AGS-sictrl cells at 8 h. In the cell viability assay, the A values of AGS-siTRAF2 cells were 51 296.00±2 631.06, 68 389.25±6 703.21 and 65 559.50±6 339.22 at 24 h, 48 h and 72 h. The values of the viability of AGS-siTRAF2 cells were significantly lower than those of AGS-sictrl cells ( P <0.001). The results of flow cytometry showed that the apoptosis rates of AGS-siTRAF2 cells were (1.42±0.07)%, (2.98±0.11)% and (1.56±0.03)% at 24 h, 48 h and 72 h, respectively, which were significantly higher than those of AGS-sictrl cells (all P <0.05). The distribution of S phase in AGS-siTRAF2 cells was (23.57±1.12)%, while that in the AGS-sictrl cells was (19.49±1.19)%. The difference was statistically significant ( P =0.012). AGS-siTRAF2 cells migrated much slower than AGS-sictrl cells from 3 h and the number of migrated AGS-sictrl cells was 121.7±6.7 while that of AGS-siTRAF2 cells was 84.0±6.6 ( P =0.002). The cell index of AGS-siTRAF2 cells was less than that of AGS-sictrl cells from 3 h. In Transwell assay, the number of invaded AGS-sictrl cells was 109.3±3.1 after 24 h of culture, significantly higher than 79.0±6.2 of AGS-siTRAF2 cells ( P =0.002). Western blot analysis showed that the expression levels of RelA, RelB, p50 and p52 in AGS-siTRAF2 cells were significantly lower than those in AGS-sictrl cells. The activities of RelA, RelB, p50 and p52 in AGS-siTRAF2 cells were 0.01±0.00, 0.01±0.01, 0.92±0.01 and 0.53±0.03, respectively, significantly lower than those of AGS-sictrl cells (all P <0.001). High TRAF2 expression (TRAF2-high) was found in 53.0% of GC samples, while TRAF2-high was only observed in 38.0% of the paired adjacent tissues ( P =0.033). The expression of TRAF2 was significantly higher in the tubular adenocarcinoma, poor differentiation advanced T, advanced N, and clinical staging ( P <0.05). The median survival time were 17 months and 78 months in the TRAF2 high-expression and low-expression groups, respectively, and the difference was statistically significant ( P =0.010). Conclusions: Depletion of TRAF2 inhibits the AGS cell growth, migration and invasion. The expression of TRAF2 is increased in gastric tumor tissue. The expression of TRAF2 is associated with the prognosis of gastric cancer.

Introduction: Non-intubated anesthesia (NIA) has been proposed for video-assisted thoracoscopic surgery (VATS), although how the benefit-to-risk of NIA compares to that of intubated general anesthesia (IGA) for certain types of patients remains unclear. Therefore, the aim of the present meta-analysis was to understand whether NIA or IGA may be more beneficial for patients undergoing VATS., Methods: A systematic search of Cochrane Library, Pubmed and Embase databases from 1968 to April 2019 was performed using predefined criteria. Studies comparing the effects of NIA or IGA for adult VATS patients were considered. The primary outcome measure was hospital stay. Pooled data were meta-analyzed using a random-effects model to determine the standard mean difference (SMD) with 95% confidence intervals (CI)., Results and Discussion: Twenty-eight studies with 2929 patients were included. The median age of participants was 56.8 years (range 21.9-76.4) and 1802 (61.5%) were male. Compared to IGA, NIA was associated with shorter hospital stay (SMD -0.57 days, 95%CI -0.78 to -0.36), lower estimated cost for hospitalization (SMD -2.83 US, 95% CI -4.33 to -1.34), shorter chest tube duration (SMD -0.32 days, 95% CI -0.47 to -0.17), and shorter postoperative fasting time (SMD, -2.76 days; 95% CI -2.98 to -2.54). NIA patients showed higher levels of total lymphocytes and natural killer cells and higher T helper/T suppressor cell ratio, but lower levels of interleukin (IL)-6, IL-8 and C-reactive protein (CRP). Moreover, NIA patients showed lower levels of fibrinogen, cortisol, procalcitonin and epinephrine., Conclusions: NIA enhances the recovery from VATS through attenuation of stress and inflammatory responses and stimulation of cellular immune function., Competing Interests: The authors have declared that no competing interests exist.

Background: Nowadays, social media are often being used by general public to create and share public messages related to their health. With the global increase in social media usage, there is a trend of posting information related to adverse drug reactions (ADR). Mining the social media data for this type of information will be helpful for pharmacological post-marketing surveillance and monitoring. Although the concept of using social media to facilitate pharmacovigilance is convincing, construction of automatic ADR detection systems remains a challenge because the corpora compiled from social media tend to be highly imbalanced, posing a major obstacle to the development of classifiers with reliable performance., Methods: Several methods have been proposed to address the challenge of imbalanced corpora. However, we are not aware of any studies that investigated the effectiveness of the strategies of dealing with the problem of imbalanced data in the context of ADR detection from social media. In light of this, we evaluated a variety of imbalanced techniques and proposed a novel word embedding-based synthetic minority over-sampling technique (WESMOTE), which synthesizes new training examples from the sentence representation based on word embeddings. We compared the performance of all methods on two large imbalanced datasets released for the purpose of detecting ADR posts., Results: In comparison with the state-of-the-art approaches, the classifiers that incorporated imbalanced classification techniques achieved comparable or better F-scores. All of our best performing configurations combined random under-sampling with techniques including the proposed WESMOTE, boosting and ensemble, implying that an integration of these approaches with under-sampling provides a reliable solution for large imbalanced social media datasets. Furthermore, ensemble-based methods like vote-based under-sampling (VUE) and random under-sampling boosting can be alternatives for the hybrid synthetic methods because both methods increase the diversity of the created weak classifiers, leading to better recall and overall F-scores for the minority classes., Conclusions: Data collected from the social media are usually very large and highly imbalanced. In order to maximize the performance of a classifier trained on such data, applications of imbalanced strategies are required. We considered several practical methods for handling imbalanced Twitter data along with their performance on the binary classification task with respect to ADRs. In conclusion, the following practical insights are gained: 1) When dealing with text classification, the proposed word embedding-based synthetic minority over-sampling technique is more effective than traditional synthetic-based over-sampling methods. 2) In cases where large amounts of training data are available, the imbalanced strategies combined with under-sampling techniques are preferred. 3) Finally, employment of advanced methods does not guarantee better performance than simpler ones such as VUE, which achieved high performance with advantages like faster building time and ease of development., (Copyright © 2019 Elsevier B.V. All rights reserved.)

Cocrystallization may alter material physicochemical properties; thus, the strategy of forming a cocrystal is generally used to improve the material performance for practical applications. In this study, two transition-metal complex cocrystals [Zn(bpy) 3 ]H 0.5 BDC·H 1.5 BDC·0.5bpy·3H 2 O ( 1 ) and [Cu 2 (BDC)(bpy) 4 ]BDC·bpy·2H 2 O ( 2 ) have been achieved using a hydrothermal reaction, where bpy and H 2 BDC represent 2,2'-bipyridine and benzene-1,3-dicarboxylic acid, respectively. Cocrystals were characterized by microanalysis, infrared spectroscopy, and UV-visible spectroscopy. Cocrystal 1 contains five components and crystallizes in a monoclinic space group P 2 1 / n . The H 0.5 BDC 1.5- , H 1.5 BDC 0.5- , and H 2 O molecules construct three-dimensional H-bonding organic framework; the [Zn(bpy) 3 ] 2+ coordination cations and uncoordinated bpy molecules reside in channels, where two coordinated bpy ligands in [Zn(bpy) 3 ] 2+ and one uncoordinated bpy adopt sandwich-type alignment via π···π stacking interactions. Cocrystal 2 with four components crystallizes in a triclinic space group P -1 to form alternating layers; the binuclear [Cu 2 (bpy) 4 (BDC)] 2+ cations and uncoordinated bpy molecules build the cationic layers, and the BDC 2- species with disordered lattice water molecules form the anionic layers. Cocrystal 1 shows intense photoluminescence at an ambient condition with a quantum yield of 14.96% and decay time of 0.48 ns, attributed to the π* → π electron transition within phenyl/pyridyl rings, and 2 exhibits magnetic behavior of an almost isolated spin system with rather weak antiferromagnetic coupling in the [Cu 2 (bpy) 4 (BDC)] 2+ cation., Competing Interests: The authors declare no competing financial interest.

Two new Pb 2+ -based coordination polymers (CPs), [Pb 2 (HBTC) 2 (DMF)] (1) and [Pb(HPTC)] (2), have been synthesized under solvothermal condition; herein, H 3 BTC and H 3 PTC represent 1,3,5-benzenetricarboxylic acid and 2,4,6-pyridine tricarboxylic acid, respectively. Both 1 and 2 were characterized by microanalysis, infrared spectra (IR), thermal gravimetric analyses (TGA), and powder X-ray diffraction (PXRD) techniques. Single-crystal structural analysis indicates that 1 and 2 show different coordination sphere around Pb 2+ ions and distinct coordination frameworks. The I 1 O 2 type three-dimensional (3D) nonporous metal-organic framework forms in 1, where the Pb 2+ ion shows holo-directed coordination geometry, while the I 0 O 2 type two-dimensional (2D) coordination polymeric layered structure forms in 2, where Pb 2+ ion shows a hemidirected coordination sphere and the 6s 2 lone electron pair in Pb 2+ ion is stereochemically active. The two CPs emit intense and long-lasting greenish phosphorescence in air at room temperature, with absolute quantum yields of 1.2% for 1 and 4.7% for 2 and decay lifetimes of 0.73 ms for 1 and 1.52 ms for 2.

Background: l-Carnitine, an abundant nutrient in red meat, accelerates atherosclerosis in mice via gut microbiota-dependent formation of trimethylamine (TMA) and trimethylamine N-oxide (TMAO) via a multistep pathway involving an atherogenic intermediate, γ-butyrobetaine (γBB). The contribution of γBB in gut microbiota-dependent l-carnitine metabolism in humans is unknown., Methods: Omnivores and vegans/vegetarians ingested deuterium-labeled l-carnitine (d3-l-carnitine) or γBB (d9-γBB), and both plasma metabolites and fecal polymicrobial transformations were examined at baseline, following oral antibiotics, or following chronic (≥2 months) l-carnitine supplementation. Human fecal commensals capable of performing each step of the l-carnitine→γBB→TMA transformation were identified., Results: Studies with oral d3-l-carnitine or d9-γBB before versus after antibiotic exposure revealed gut microbiota contribution to the initial 2 steps in a metaorganismal l-carnitine→γBB→TMA→TMAO pathway in subjects. Moreover, a striking increase in d3-TMAO generation was observed in omnivores over vegans/vegetarians (>20-fold; P = 0.001) following oral d3-l-carnitine ingestion, whereas fasting endogenous plasma l-carnitine and γBB levels were similar in vegans/vegetarians (n = 32) versus omnivores (n = 40). Fecal metabolic transformation studies, and oral isotope tracer studies before versus after chronic l-carnitine supplementation, revealed that omnivores and vegans/vegetarians alike rapidly converted carnitine to γBB, whereas the second gut microbial transformation, γBB→TMA, was diet inducible (l-carnitine, omnivorous). Extensive anaerobic subculturing of human feces identified no single commensal capable of l-carnitine→TMA transformation, multiple community members that converted l-carnitine to γBB, and only 1 Clostridiales bacterium, Emergencia timonensis, that converted γBB to TMA. In coculture, E. timonensis promoted the complete l-carnitine→TMA transformation., Conclusion: In humans, dietary l-carnitine is converted into the atherosclerosis- and thrombosis-promoting metabolite TMAO via 2 sequential gut microbiota-dependent transformations: (a) initial rapid generation of the atherogenic intermediate γBB, followed by (b) transformation into TMA via low-abundance microbiota in omnivores, and to a markedly lower extent, in vegans/vegetarians. Gut microbiota γBB→TMA/TMAO transformation is induced by omnivorous dietary patterns and chronic l-carnitine exposure., Trial Registration: ClinicalTrials.gov NCT01731236., Funding: NIH and Office of Dietary Supplements grants HL103866, HL126827, and DK106000, and the Leducq Foundation.

The detection of MicroRNA (miRNA) mentions in scientific literature facilitates researchers with the ability to find relevant and appropriate literature based on queries formulated using miRNA information. Considering most published biological studies elaborated on signal transduction pathways or genetic regulatory information in the form of figure captions, the extraction of miRNA from both the main content and figure captions of a manuscript is useful in aggregate analysis and comparative analysis of the studies published. In this study, we present a statistical principle-based miRNA recognition and normalization method to identify miRNAs and link them to the identifiers in the Rfam database. As one of the core components in the text mining pipeline of the database miRTarBase, the proposed method combined the advantages of previous works relying on pattern, dictionary and supervised learning and provided an integrated solution for the problem of miRNA identification. Furthermore, the knowledge learned from the training data was organized in a human-interpretable manner to understand the reason why the system considers a span of text as a miRNA mention, and the represented knowledge can be further complemented by domain experts. We studied the ambiguity level of miRNA nomenclature to connect the miRNA mentions to the Rfam database and evaluated the performance of our approach on two datasets: the BioCreative VI Bio-ID corpus and the miRNA interaction corpus by extending the later corpus with additional Rfam normalization information. Our study highlights and also proposes a better understanding of the challenges associated with miRNA identification and normalization in scientific literature and the research gap that needs to be further explored in prospective studies., (© The Author(s) 2019. Published by Oxford University Press.)

The Precision Medicine Initiative is a multicenter effort aiming at formulating personalized treatments leveraging on individual patient data (clinical, genome sequence and functional genomic data) together with the information in large knowledge bases (KBs) that integrate genome annotation, disease association studies, electronic health records and other data types. The biomedical literature provides a rich foundation for populating these KBs, reporting genetic and molecular interactions that provide the scaffold for the cellular regulatory systems and detailing the influence of genetic variants in these interactions. The goal of BioCreative VI Precision Medicine Track was to extract this particular type of information and was organized in two tasks: (i) document triage task, focused on identifying scientific literature containing experimentally verified protein-protein interactions (PPIs) affected by genetic mutations and (ii) relation extraction task, focused on extracting the affected interactions (protein pairs). To assist system developers and task participants, a large-scale corpus of PubMed documents was manually annotated for this task. Ten teams worldwide contributed 22 distinct text-mining models for the document triage task, and six teams worldwide contributed 14 different text-mining systems for the relation extraction task. When comparing the text-mining system predictions with human annotations, for the triage task, the best F-score was 69.06%, the best precision was 62.89%, the best recall was 98.0% and the best average precision was 72.5%. For the relation extraction task, when taking homologous genes into account, the best F-score was 37.73%, the best precision was 46.5% and the best recall was 54.1%. Submitted systems explored a wide range of methods, from traditional rule-based, statistical and machine learning systems to state-of-the-art deep learning methods. Given the level of participation and the individual team results we find the precision medicine track to be successful in engaging the text-mining research community. In the meantime, the track produced a manually annotated corpus of 5509 PubMed documents developed by BioGRID curators and relevant for precision medicine. The data set is freely available to the community, and the specific interactions have been integrated into the BioGRID data set. In addition, this challenge provided the first results of automatically identifying PubMed articles that describe PPI affected by mutations, as well as extracting the affected relations from those articles. Still, much progress is needed for computer-assisted precision medicine text mining to become mainstream. Future work should focus on addressing the remaining technical challenges and incorporating the practical benefits of text-mining tools into real-world precision medicine information-related curation.

Background and Objective: Efficiently capturing the severity of positive valence symptoms could aid in risk stratification for adverse outcomes among patients with psychiatric disorders and identify optimal treatment strategies for patient subgroups. Motivated by the success of convolutional neural networks (CNNs) in classification tasks, we studied the application of various CNN architectures and their performance in predicting the severity of positive valence symptoms in patients with psychiatric disorders based on initial psychiatric evaluation records., Methods: Psychiatric evaluation records contain unstructured text and semi-structured data such as question-answer pairs. For a given record, we tokenise and normalise the semi-structured content. Pre-processed tokenised words are represented as one-hot encoded word vectors. We then apply different configurations of convolutional and max pooling layers to automatically learn important features from various word representations. We conducted a series of experiments to explore the effect of different CNN architectures on the classification of psychiatric records., Results: Our best CNN model achieved a mean absolute error (MAE) of 0.539 and a normalized MAE of 0.785 on the test dataset, which is comparable to the other well-known text classification algorithms studied in this work. Our results also suggest that the normalisation step has a great impact on the performance of the developed models., Conclusions: We demonstrate that normalisation of the semi-structured contents can improve the MAE among all CNN configurations. Without advanced feature engineering, CNN-based approaches can provide a comparable solution for classifying positive valence symptom severity in initial psychiatric evaluation records. Although word embedding is well known for its ability to capture relatively low-dimensional similarity between words, our experimental results show that pre-trained embeddings do not improve the classification performance. This phenomenon may be due to the inability of word embeddings to capture problem specific contextual semantic information implying the quality of the employing embedding is critical for obtaining an accurate CNN model., Competing Interests: The authors have declared that no competing interests exist.

Objective: This study aimed to determine whether high tidal volume (HTV) induce mitochondria damage and mitophagy, contributing to the release of mitochondrial DNA (mtDNA). Another aim of the present study was to investigate the role and mechanism of mtDNA in ventilator-induced lung injury (VILI) in rats., Methods: Rats were tracheotomized and allowed to breathe spontaneously or mechanically ventilated for 4 h. After that, lung injury was assessed. Inhibition of toll-like receptor 9 (TLR9), named ODN2088, was used to determine the involvement of TLR9/myeloid differentiation factor 88 (MyD88)/nuclear factor-κB (NF-κB) signaling pathway in VILI. The mitochondrial damage and release of mtDNA were assessed. Pharmacological inhibition of mtDNA (chloroquine) was used to determine whether mtDNA trigger inflammation via TLR9 in VILI. EDU-labeled mtDNA deriving from mitophagy was assessed by immunofluorescence. The role of mitophagy in VILI was shown by administration of antimycin A and cyclosporine A., Main Results: Rats subjected to HTV showed more severe pulmonary edema and inflammation than the other rats. The decreased expression of TLR9, MyD88, and NF-κB were observed following the use of ODN2088. Mechanical ventilation (MV) with HTV damaged mitochondria which resulted in dysfunctional ATP synthesis, accumulation of reactive oxygen species, and loss of mitochondrial membrane potential. Moreover, the results of distribution of fluorescence in rats upon HTV stimulation indicated that mtDNA cleavage was associated with mitophagy. The expression levels of mitophagy related genes (LC3B-II/LC3B-I, PINK1, Parkin, and mitofusin 1) in animals ventilated with HTV were significantly upregulated. Administration of antimycin A aggregated the histological changes and inflammation after MV, but these effects were attenuated when administered in the presence of cyclosporine A., Conclusion: MV with HTV induces mitochondrial damage and mitophagy, contributing to the release of mtDNA, which may be induced VILI in rat via TLR9/MyD88/NF-κB signaling pathway.

Immunotherapeutic agents have demonstrated encouraging signs of clinical utility in non-Hodgkin lymphoma. The goal of this study is to analyze the immune characteristics of Chinese patients with diffuse large B-cell lymphoma (DLBCL) to inform the development of immunotherapies in this patient population. Tumor samples from 211 DLBCL patients were analyzed for cell of origin (COO) and immune characteristics using the NanoString platform as well as MYC protein expression through immunohistochemistry. Lower incidence of the germinal center B-cell (GCB) subtype (93/211, 44.1%) was observed in this cohort. Compared to the GCB subtype, the activated B-cell (ABC) subtype was associated with significantly increased expression of multiple pro-inflammatory gene signatures and decreased expression of anti-inflammatory gene signatures. Instead of affecting the pro-inflammatory genes, MYC protein overexpression showed a negative correlation with the expression of T-cell receptor (TCR) and T regulatory genes as well as the OX40 gene. Regardless of COO, higher PD-L1 or IDO1 gene expression correlated with increased expression of T effector and Interferon-γ gene signatures while the expression of multiple oncogenes including ACTR3B, ERBB2, AKT2 and SMARCD1 was down-regulated. Our findings may thus be helpful in guiding further development of immunotherapies for the different subsets of Chinese DLBCL patients., (Copyright © 2018 The Authors. Published by Elsevier B.V. All rights reserved.)

Carboxyl terminus of Hsc-70-interacting protein (CHIP) is an E3 ubiquitin ligase that induces the ubiquitination and degradation of numerous tumor-associated proteins and serves as a suppressor or promoter in tumor progression. To date, the molecular mechanism of CHIP in prostate cancer remains unknown. Therefore, the present study investigated the biological function of CHIP in prostate cancer cells and obtained evidence that CHIP expression is upregulated in prostate cancer tissues. The CHIP vector was introduced into DU145 cancer cells and the cell biological behaviour was examined through a series of experiments, including cell growth, cell apoptosis and migration and invasion assays. The results indicated that the overexpression of CHIP in DU145 prostatic cancer cells promoted cell proliferation through activation of the protein kinase B (AKT) signaling pathway, which subsequently increased cyclin D1 protein levels and decreased p21 and p27 protein levels. The overexpression of CHIP significantly increased the migration and invasion of the DU145 cells, which is possible due to activation of the AKT signaling pathway and upregulation of vimentin. The expression level of CHIP was observed to be increased in human prostate cancer tissues compared with the adjacent normal tissue. Furthermore, the CHIP expression level exhibited a positively association with the Gleason score of the patents. These findings indicate that CHIP functions as an oncogene in prostate cancer.

Claudin18 (CLDN18) is necessary for intercellular junctions and is reported to be involved in cell migration and metastasis, making it like an oncogene in various cancer types. However, the biological function and regulatory mechanisms of CLDN18 in lung adenocarcinoma are not yet clear. In this study, we found downregulation of miR-767-3p and upregulation of CLDN18 in lung adenocarcinoma tissue and cell lines. In addition, there was a negative correlation between the expression of miR-767-3p and CLDN18 in lung adenocarcinoma. Double luciferase reporter gene analysis showed that miR-767-3p modulates the expression of CLDN18 by binding its 3'-untranslated regions (3'-UTR). Knockdown of CLDN18 results in a decrease in the growth, migration, and invasion of lung adenocarcinoma cells. Although overexpression of miR-767-3p inhibits lung adenocarcinoma cell growth and migration, these effects can be rescued by reexpressing CLDN18. In summary, the data suggest that miR-767-3p inhibits tumor cell proliferation, migration, and invasion by targeting CLDN18, providing a promising therapeutic target for lung adenocarcinoma.

Background: Esophageal squamous cell carcinoma (ESCC) is often diagnosed at an advanced and incurable stage. Information on driver genes and prognosticators in ESCC remains incomplete. The objective was to elucidate significantly mutated genes (SMGs), mutational signatures, and prognosticators in ESCC., Patients and Methods: Three MutSig algorithms (i.e. MutSigCV, MutSigCL and MutSigFN) and '20/20+' ratio-metric were employed to identify SMGs. Nonnegative matrix factorization was used to decipher mutational signatures. Kaplan-Meier survival analysis, multivariate Cox and logistic regression models were applied to analyze association between mutational features and clinical parameters., Results: We identified 26 SMGs, including 8 novel (NAV3, TENM3, PTCH1, TGFBR2, RIPK4, PBRM1, USP8 and BAP1) and 18 that have been previously reported. Three mutational signatures were identified to be prevalent in ESCC including clocklike C>T at CpG, APOBEC overactive C>T at TpCp[A/T], and a signature featured by T>C substitution. The T>C mutational signature was significantly correlated with alcohol consumption (OR: 3.59; 95% CI: 2.30-5.67; P < 0.001). This alcohol consumption signature was also observed in liver cancer and head and neck squamous cell carcinoma, and its mutational activity was substantially higher in samples with mutations in TP53. Survival analysis revealed that TENM3 mutations (HR: 5.54; CI: 2.68-11.45; P < 0.001) and TP53 hotspot mutation p.R213* (HR: 3.37; CI: 1.73-8.06; P < 0.001) were significantly associated with shortened survival outcome. The association remained statistically significant after controlling for age, gender, TNM stage and tumor grade., Conclusions: We have uncovered several new SMGs in ESCC and defined an alcohol consumption related mutational signature. TENM3 mutations and the TP53 hotspot mutation p.R213* are independent prognosticators for poor survival in ESCC.

The aim of this study is to compare the effect of conscious sedation (CS) with general anesthesia (GA) on clinical outcomes in patients with acute ischemic stroke (AIS) undergoing endovascular therapy (EVT). MEDLINE, EMBASE, and Cochrane Central Registers of Controlled Trials (from inception to July 2017) were searched for reports on CS and GA of AIS undergoing EVT. Two reviewers assessed the eligibility of the identified studies and extracted data. Data were analyzed using the fixed-effects model, and the sources of heterogeneity were explored by sensitive analysis. Trial sequential analysis was conducted to monitor boundaries for the limitation of global type I error, and GRADE system was demonstrated to evaluate the quality of evidence. A total of thirteen studies were finally identified. Pooled analysis of the incidence of mRS score ≦ 2 after hospital discharge and one or three months in the CS group was higher than that in the GA group. The all-causing mortality of AIS patients in the CS group was lower than that in the GA group. There were no differences in the proportion of IA rtPA and thrombolysis between the two groups. Compared with AIS patients receiving GA, the all-causing mortality in the AIS patients receiving CS was decreased, while incidence of mRS score ≦ 2 at hospital discharge and one or three months was increased.

The goal of this study is to evaluate PD-L1 prevalence and its association with major clinical characteristics in Chinese non-small cell lung cancer (NSCLC) patients to inform the clinical development of anti-PD1/PD-L1 agents in this population. We used phosphatase and tensin homolog (PTEN) expression through IHC as a surrogate tissue quality marker to screen surgical NSCLC samples in tissue microarray (TMA; 172 cases) or whole-section (268 cases) format. The samples were then analyzed with a clinically validated PD-L1 IHC assay. The results were correlated with baseline characteristics and clinical outcomes. PTEN IHC showed that 108 TMA samples and 105 whole-section samples qualified for PD-L1 IHC. With a clinically relevant cutoff, 41.7% of the TMA samples were PD-L1 positive. PD-L1 level was much lower in EGFR-mutant patients and seemed to be a favorable prognostic factor for both overall survival (OS) and recurrence-free survival (RFS). These findings were confirmed in the whole-section samples except that their survival data were not mature enough for correlation analysis. In summary, PD-L1 expression was detected in approximately 40% of PTEN-qualified Chinese NSCLC samples, negatively correlated with EGFR mutation and seemed to be a favorable prognostic factor for both OS and RFS. Notably, the different results from PTEN-qualified and PTEN-disqualified samples underscore the importance of tissue quality control prior to biomarker testing.

Recent advances in biological research reveal that the majority of the experiments strive for comprehensive exploration of the biological system rather than targeting specific biological entities. The qualitative and quantitative findings of the investigations are often exclusively available in the form of figures in published papers. There is no denying that such findings have been instrumental in intensive understanding of biological processes and pathways. However, data as such is unacknowledged by machines as the descriptions in the figure captions comprise of sumptuous information in an ambiguous manner. The abbreviated term 'SIN' exemplifies such issue as it may stand for Sindbis virus or the sex-lethal interactor gene (Drosophila melanogaster). To overcome this ambiguity, entities should be identified by linking them to the respective entries in notable biological databases. Among all entity types, the task of identifying species plays a pivotal role in disambiguating related entities in the text. In this study, we present our species identification tool SPRENO (Species Recognition and Normalization), which is established for recognizing organism terms mentioned in figure captions and linking them to the NCBI taxonomy database by exploiting the contextual information from both the figure caption and the corresponding full text. To determine the ID of ambiguous organism mentions, two disambiguation methods have been developed. One is based on the majority rule to select the ID that has been successfully linked to previously mentioned organism terms. The other is a convolutional neural network (CNN) model trained by learning both the context and the distance information of the target organism mention. As a system based on the majority rule, SPRENO was one of the top-ranked systems in the BioCreative VI BioID track and achieved micro F-scores of 0.776 (entity recognition) and 0.755 (entity normalization) on the official test set, respectively. Additionally, the SPRENO-CNN exhibited better precisions with lower recalls and F-scores (0.720/0.711 for entity recognition/normalization). SPRENO is freely available at https://bigodatamining.github.io/software/201801/.Database URL: https://bigodatamining.github.io/software/201801/.

Evidence has revealed interesting associations of clinical and social parameters with violent behaviors of patients with psychiatric disorders. Men are more violent preceding and during hospitalization, whereas women are more violent than men throughout the 3days following a hospital admission. It has also been proven that mental disorders may be a consistent risk factor for the occurrence of violence. In order to better understand violent behaviors of patients with psychiatric disorders, it is important to investigate both the clinical symptoms and psychosocial factors that accompany violence in these patients. In this study, we utilized a dataset released by the Partners Healthcare and Neuropsychiatric Genome-scale and RDoC Individualized Domains project of Harvard Medical School to develop a unique text mining pipeline that processes unstructured clinical data in order to recognize clinical and social parameters such asage, gender, history of alcohol use, and violent behaviors, and explored the associations between these parameters and violent behaviors of patients with psychiatric disorders. The aim of our work was to demonstrate the feasibility of mining factors that are strongly associated with violent behaviors among psychiatric patients from unstructured psychiatric evaluation records using clinical text mining. Experiment results showed that stimulants, followed by a family history of violent behavior, suicidal behaviors, and financial stress were strongly associated with violent behaviors. Key aspects explicated in this paper include employing our text mining pipeline to extract clinical and social factors linked with violent behaviors, generating association rules to uncover possible associations between these factors and violent behaviors, and lastly the ranking of top rules associated with violent behaviors using statistical analysis and interpretation., (Copyright © 2017. Published by Elsevier Inc.)

Genome-wide association studies have identified 21 susceptibility loci associated with melanoma. These loci implicate genes affecting pigmentation, nevus count, telomere maintenance, and DNA repair in melanoma risk. Here, we report the results of a two-stage genome-wide association study of melanoma. The stage 1 discovery phase consisted of 4,842 self-reported melanoma cases and 286,565 controls of European ancestry from the 23andMe research cohort and the stage 2 replication phase consisted of 1,804 melanoma cases and 1,026 controls from the University of Texas M.D. Anderson Cancer Center. We performed a combined meta-analysis totaling 6,628 melanoma cases and 287,591 controls. Our study replicates 20 of 21 previously known melanoma-loci and confirms the association of the telomerase reverse transcriptase, TERT, with melanoma susceptibility at genome-wide significance. In addition, we uncover a novel polymorphism, rs187843643 (OR = 1.96; 95% CI = [1.54, 2.48]; P = 3.53 x 10-8), associated with melanoma. The SNP rs187842643 lies within a noncoding RNA 177kb downstream of BASP1 (brain associated protein-1). We find that BASP1 expression is suppressed in melanoma as compared with benign nevi, providing additional evidence for a putative role in melanoma pathogenesis.

Hepatocellular carcinoma (HCC), one of the most common causes of cancer-related deaths, carries a 5-year survival rate of 18%, underscoring the need for robust biomarkers. In spite of the increased availability of HCC related literatures, many of the promising biomarkers reported have not been validated for clinical use. To narrow down the wide range of possible biomarkers for further clinical validation, bioinformaticians need to sort them out using information provided in published works. Biomedical text mining is an automated way to obtain information of interest within the massive collection of biomedical knowledge, thus enabling extraction of data for biomarkers associated with certain diseases. This method can significantly reduce both the time and effort spent on studying important maladies such as liver diseases. Herein, we report a text mining-aided curation pipeline to identify potential biomarkers for liver cancer. The curation pipeline integrates PubMed E-Utilities to collect abstracts from PubMed and recognize several types of named entities by machine learning-based and pattern-based methods. Genes/proteins from evidential sentences were classified as candidate biomarkers using a convolutional neural network. Lastly, extracted biomarkers were ranked depending on several criteria, such as the frequency of keywords and articles and the journal impact factor, and then integrated into a meaningful list for bioinformaticians. Based on the developed pipeline, we constructed MarkerHub, which contains 2128 candidate biomarkers extracted from PubMed publications from 2008 to 2017. Database URL: http://markerhub.iis.sinica.edu.tw., (© The Author(s) 2017. Published by Oxford University Press.)