Search

Abstract Patient recruitment is challenging for clinical trials. We introduce TrialGPT, an end-to-end framework for zero-shot patient-to-trial matching with large language models. TrialGPT comprises three modules: it first performs large-scale filtering to retrieve candidate trials (TrialGPT-Retrieval); then predicts criterion-level patient eligibility (TrialGPT-Matching); and finally generates trial-level scores (TrialGPT-Ranking). We evaluate TrialGPT on three cohorts of 183 synthetic patients with over 75,000 trial annotations. TrialGPT-Retrieval can recall over 90% of relevant trials using less than 6% of the initial collection. Manual evaluations on 1015 patient-criterion pairs show that TrialGPT-Matching achieves an accuracy of 87.3% with faithful explanations, close to the expert performance. The TrialGPT-Ranking scores are highly correlated with human judgments and outperform the best-competing models by 43.8% in ranking and excluding trials. Furthermore, our user study reveals that TrialGPT can reduce the screening time by 42.6% in patient recruitment. Overall, these results have demonstrated promising opportunities for patient-to-trial matching with TrialGPT.

Abstract Many tomb murals have punctate losses, cracks, and craquelure due to underground subsidence and changes in their physical support. Visual non-destructive detection techniques enable rapid assessment of how much tomb murals are affected by cracking, providing suggestions for their conservation. However, tomb murals are typically created by sketching outlines and then colored. Detailed sketches can easily interfere with crack detection work, requiring the use of deep learning network to better learn crack features and improve detection accuracy. At the same time the limited data of tomb mural presents a challenge to build a deep learning network. To address these issues, this paper introduces a novel dual-attention detection network (DADNet) for crack segmentation of tomb murals. In this work, a customized dataset is first constructed by collecting mural images from the Tang Dynasty tombs. Then the ConvNeXt framework serves as the basis for feature extraction, enhancing the process. Lastly, a dual-attention module utilizing neighborhood attention and biaxial attention is employed to accurately identify the crack regions. Neighborhood attention performs a local self-attention operation around the pixel point, addressing the limitations of self-attention. This approach significantly reduces computational demands as the image size increases. Biaxial attention performs attention calculations in the horizontal and vertical directions. This compensates for the limitation of neighborhood attention in capturing global dependencies. Our DADNet outperformed the competing methods, achieving the highest recorded scores of 78.95% for MIoU and 61.05% for the Jaccard index.

Objective: To propose Deep-RPD-Net, a 3-dimensional deep learning network with semisupervised learning (SSL) for the detection of reticular pseudodrusen (RPD) on spectral-domain OCT scans, explain its decision-making, and compare it with baseline methods. Design: Deep learning model development. Participants: Three hundred fifteen participants from the Age-Related Eye Disease Study 2 Ancillary OCT Study (AREDS2) and 161 participants from the Dark Adaptation in Age-related Macular Degeneration Study (DAAMD). Methods: Two datasets comprising of 1304 (826 labeled) and 1479 (1366 labeled) OCT scans were used to develop and evaluate Deep-RPD-Net and baseline models. The AREDS2 RPD labels were transferred from fundus autofluorescence images, which were captured at the same visit for OCT scans, and DAAMD RPD labels were obtained from the Wisconsin reading center. The datasets were divided into 70%, 10%, and 20% at the participant level for training, validation, and test sets, respectively, for the baseline model. Then, SSL was used with the unlabeled OCT scans to improve the trained model. The performance of Deep-RPD-Net was compared to that of 3 retina specialists on a subset of 50 OCT scans for each dataset. En face and B-scan heatmaps of all networks were visualized and graded on 25 OCT scans with positive labels, using a scale of 1 to 4, to explore the models' decision-making. Main Outcome Measures: Accuracy and area under the receiver-operating characteristic curve (AUROC). Results: Deep-RPD-Net achieved the highest performance metrics, with accuracy and AUROC of 0.81 (95% confidence interval [CI]: 0.76–0.87) and 0.91 (95% CI: 0.86–0.95) on the AREDS2 dataset and 0.80 (95% CI: 0.75–0.84) and 0.86 (95% CI: 0.79–0.91) on the DAAMD dataset. On the subjective test, it achieved accuracy of 0.84 compared with 0.76 for the most accurate retina specialist on the AREDS2 dataset and 0.82 compared with 0.84 on the DAAMD dataset. It also achieved the highest visualization grades, of 3.26 and 3.32 for en face and B-scan heatmaps, respectively. Conclusions: Deep-RPD-Net was able to detect RPD accurately from OCT scans. The visualizations of Deep-RPD-Net were the most explainable to the retina specialist with the highest accuracy. The code and pretrained models are publicly available at https://github.com/ncbi-nlp/Deep-RPD-Net. Financial Disclosure(s): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Abstract Expert curation is essential to capture knowledge of enzyme functions from the scientific literature in FAIR open knowledgebases but cannot keep pace with the rate of new discoveries and new publications. In this work we present EnzChemRED, for Enzyme Chemistry Relation Extraction Dataset, a new training and benchmarking dataset to support the development of Natural Language Processing (NLP) methods such as (large) language models that can assist enzyme curation. EnzChemRED consists of 1,210 expert curated PubMed abstracts where enzymes and the chemical reactions they catalyze are annotated using identifiers from the protein knowledgebase UniProtKB and the chemical ontology ChEBI. We show that fine-tuning language models with EnzChemRED significantly boosts their ability to identify proteins and chemicals in text (86.30% F1 score) and to extract the chemical conversions (86.66% F1 score) and the enzymes that catalyze those conversions (83.79% F1 score). We apply our methods to abstracts at PubMed scale to create a draft map of enzyme functions in literature to guide curation efforts in UniProtKB and the reaction knowledgebase Rhea.

Abstract Background Large language models like GPT-3.5-turbo and GPT-4 hold promise for healthcare professionals, but they may inadvertently inherit biases during their training, potentially affecting their utility in medical applications. Despite few attempts in the past, the precise impact and extent of these biases remain uncertain. Methods We use LLMs to generate responses that predict hospitalization, cost and mortality based on real patient cases. We manually examine the generated responses to identify biases. Results We find that these models tend to project higher costs and longer hospitalizations for white populations and exhibit optimistic views in challenging medical scenarios with much higher survival rates. These biases, which mirror real-world healthcare disparities, are evident in the generation of patient backgrounds, the association of specific diseases with certain racial and ethnic groups, and disparities in treatment recommendations, etc. Conclusions Our findings underscore the critical need for future research to address and mitigate biases in language models, especially in critical healthcare applications, to ensure fair and accurate outcomes for all patients.

Abstract Large language models (LLMs) hold great promise in summarizing medical evidence. Most recent studies focus on the application of proprietary LLMs. Using proprietary LLMs introduces multiple risk factors, including a lack of transparency and vendor dependency. While open-source LLMs allow better transparency and customization, their performance falls short compared to the proprietary ones. In this study, we investigated to what extent fine-tuning open-source LLMs can further improve their performance. Utilizing a benchmark dataset, MedReview, consisting of 8161 pairs of systematic reviews and summaries, we fine-tuned three broadly-used, open-sourced LLMs, namely PRIMERA, LongT5, and Llama-2. Overall, the performance of open-source models was all improved after fine-tuning. The performance of fine-tuned LongT5 is close to GPT-3.5 with zero-shot settings. Furthermore, smaller fine-tuned models sometimes even demonstrated superior performance compared to larger zero-shot models. The above trends of improvement were manifested in both a human evaluation and a larger-scale GPT4-simulated evaluation.

Objective To observe the changes of coagulation factor activity and protein content of the thawed fresh frozen plasma (FFP) and fresh liquid plasma (FLP) during storage at 2-6℃, and to provide reference for exploring the appropriate storage time of FFP at 2-6℃ after thawing. Methods The small-thaw group and the large-thaw group were respectively detected for the activity of coagulation factor FⅤ (FⅤ∶C) and FⅧ(FⅧ∶C), and the levels of fibrinogen (Fib), total protein (TP) and albumin (Alb) in TTP at 1, 2, 3, 4, 5, 6 and 7 days after thawing. And the FLP was detected for FⅤ∶C, FⅧ∶C, Fib, TP and Alb at 1, 2, 3, 4, 5, 6, 7 days and 1, 6, 11, 16, 21, 26 and 31 days after preparation, respectively. Results In FFP group, FⅧ∶C decreased gradually with the prolongation of storage time after melting (P0.05). In FLP group, FⅧ∶C decreased gradually with the prolongation of storage time after melting (P0.05), but it decreased gradually in 31-day storage group (P0.05). Conclusion Although FⅧ∶C is decreased in thawed FFP stored at 2-6℃ for 7 days, it is still about 52.4%, which should be able to play a normal role in clinical practice.

Abstract Deep learning has enabled breakthroughs in automated diagnosis from medical imaging, with many successful applications in ophthalmology. However, standard medical image classification approaches only assess disease presence at the time of acquisition, neglecting the common clinical setting of longitudinal imaging. For slow, progressive eye diseases like age-related macular degeneration (AMD) and primary open-angle glaucoma (POAG), patients undergo repeated imaging over time to track disease progression and forecasting the future risk of developing a disease is critical to properly plan treatment. Our proposed Longitudinal Transformer for Survival Analysis (LTSA) enables dynamic disease prognosis from longitudinal medical imaging, modeling the time to disease from sequences of fundus photography images captured over long, irregular time periods. Using longitudinal imaging data from the Age-Related Eye Disease Study (AREDS) and Ocular Hypertension Treatment Study (OHTS), LTSA significantly outperformed a single-image baseline in 19/20 head-to-head comparisons on late AMD prognosis and 18/20 comparisons on POAG prognosis. A temporal attention analysis also suggested that, while the most recent image is typically the most influential, prior imaging still provides additional prognostic value.

Abstract Recent studies indicate that Generative Pre-trained Transformer 4 with Vision (GPT-4V) outperforms human physicians in medical challenge tasks. However, these evaluations primarily focused on the accuracy of multi-choice questions alone. Our study extends the current scope by conducting a comprehensive analysis of GPT-4V’s rationales of image comprehension, recall of medical knowledge, and step-by-step multimodal reasoning when solving New England Journal of Medicine (NEJM) Image Challenges—an imaging quiz designed to test the knowledge and diagnostic capabilities of medical professionals. Evaluation results confirmed that GPT-4V performs comparatively to human physicians regarding multi-choice accuracy (81.6% vs. 77.8%). GPT-4V also performs well in cases where physicians incorrectly answer, with over 78% accuracy. However, we discovered that GPT-4V frequently presents flawed rationales in cases where it makes the correct final choices (35.5%), most prominent in image comprehension (27.2%). Regardless of GPT-4V’s high accuracy in multi-choice questions, our findings emphasize the necessity for further in-depth evaluations of its rationales before integrating such multimodal AI models into clinical workflows.

BackgroundMedical texts present significant domain-specific challenges, and manually curating these texts is a time-consuming and labor-intensive process. To address this, natural language processing (NLP) algorithms have been developed to automate text processing. In the biomedical field, various toolkits for text processing exist, which have greatly improved the efficiency of handling unstructured text. However, these existing toolkits tend to emphasize different perspectives, and none of them offer generation capabilities, leaving a significant gap in the current offerings. ObjectiveThis study aims to describe the development and preliminary evaluation of Ascle. Ascle is tailored for biomedical researchers and clinical staff with an easy-to-use, all-in-one solution that requires minimal programming expertise. For the first time, Ascle provides 4 advanced and challenging generative functions: question-answering, text summarization, text simplification, and machine translation. In addition, Ascle integrates 12 essential NLP functions, along with query and search capabilities for clinical databases. MethodsWe fine-tuned 32 domain-specific language models and evaluated them thoroughly on 27 established benchmarks. In addition, for the question-answering task, we developed a retrieval-augmented generation (RAG) framework for large language models that incorporated a medical knowledge graph with ranking techniques to enhance the reliability of generated answers. Additionally, we conducted a physician validation to assess the quality of generated content beyond automated metrics. ResultsThe fine-tuned models and RAG framework consistently enhanced text generation tasks. For example, the fine-tuned models improved the machine translation task by 20.27 in terms of BLEU score. In the question-answering task, the RAG framework raised the ROUGE-L score by 18% over the vanilla models. Physician validation of generated answers showed high scores for readability (4.95/5) and relevancy (4.43/5), with a lower score for accuracy (3.90/5) and completeness (3.31/5). ConclusionsThis study introduces the development and evaluation of Ascle, a user-friendly NLP toolkit designed for medical text generation. All code is publicly available through the Ascle GitHub repository. All fine-tuned language models can be accessed through Hugging Face.

Abstract Removing unwanted materials, such as organic coatings and soil, from the cultural relic surface is a complex and significant task in the field of cultural heritage conservation. Microemulsion-loaded gel can effectively and safely remove those organic coatings and soil. Here, we employed a simple solvent exchange strategy to prepare a microemulsion-loaded polyvinyl alcohol/polyethyleneimine (PVA/PEI) hydrogel. First, PVA and PEI were dissolved into DMSO to form a gel. Then, the gel was immersed into a microemulsion composed of water, ethyl acetate, propylene carbonate, sodium dodecyl sulfate, and 1-pentanol to exchange DMSO. Microemulsion-loaded PVA/PEI hydrogel can be synthesized by completely substituting DMSO. To investigate the microstructure, rheological properties, and mechanical properties of the gel, scanning electron microscopy, a rheometer, and a universal testing machine were used, respectively. Fourier transform infrared (FT-IR) analysis was conducted to explore the synthesis mechanism and confirm the successful loading of microemulsion within the microemulsion-loaded PVA/PEI hydrogel. Furthermore, FT-IR, a depth-of-field microscope, and a glossmeter were utilized to evaluate the cleaning efficiency of the microemulsion-loaded PVA/PEI hydrogel for removing animal glue and soil from the surfaces of cultural relics. Moreover, an X-ray fluorescence spectrometer was used to analyze the element component of the ancient coin. The application results showed that the microemulsion-loaded PVA/PEI hydrogel can effectively remove animal glue from an ancient wall painting surface. Moreover, it is capable of removing soil from an ancient coin surface as well, which helped to confirm the age of the coin. This offers a novel method to prepare microemulsion-loaded hydrogel and demonstrates great potential in the cleaning for cultural heritage.

High-accuracy and high-precision inter-satellite ranging enhances the orbital accuracy of the Global Navigation Satellite System and facilitates Autonomous Navigation without requiring ground stations. This study proposes a novel phase discrimination method based on pseudo-random code phase modulation coherent laser ranging, which solves the problem of mutual restriction between ranging accuracy and ranging precision in the traditional method. The early–late correlation peaks are obtained via data shifting, while the early and late codes remain unchanged. The characteristic parameters of the early–late discriminator model are calibrated by the actual ranging system, which achieves enhanced ranging accuracy and precision simultaneously. Ground test results indicate that for the static target, the accuracy of the distance measurement is 0.56 mm, while the precision is 0.34 mm. The ranging accuracy of the proposed method has improved by a factor of 91 compared to the traditional method. For dynamic targets, the accuracies of the distance and speed measurements are 0.38 mm and 0.44 mm/s, respectively.

Abstract Background Cataract diagnosis typically requires in-person evaluation by an ophthalmologist. However, color fundus photography (CFP) is widely performed outside ophthalmology clinics, which could be exploited to increase the accessibility of cataract screening by automated detection. Methods DeepOpacityNet was developed to detect cataracts from CFP and highlight the most relevant CFP features associated with cataracts. We used 17,514 CFPs from 2573 AREDS2 participants curated from the Age-Related Eye Diseases Study 2 (AREDS2) dataset, of which 8681 CFPs were labeled with cataracts. The ground truth labels were transferred from slit-lamp examination of nuclear cataracts and reading center grading of anterior segment photographs for cortical and posterior subcapsular cataracts. DeepOpacityNet was internally validated on an independent test set (20%), compared to three ophthalmologists on a subset of the test set (100 CFPs), externally validated on three datasets obtained from the Singapore Epidemiology of Eye Diseases study (SEED), and visualized to highlight important features. Results Internally, DeepOpacityNet achieved a superior accuracy of 0.66 (95% confidence interval (CI): 0.64–0.68) and an area under the curve (AUC) of 0.72 (95% CI: 0.70–0.74), compared to that of other state-of-the-art methods. DeepOpacityNet achieved an accuracy of 0.75, compared to an accuracy of 0.67 for the ophthalmologist with the highest performance. Externally, DeepOpacityNet achieved AUC scores of 0.86, 0.88, and 0.89 on SEED datasets, demonstrating the generalizability of our proposed method. Visualizations show that the visibility of blood vessels could be characteristic of cataract absence while blurred regions could be characteristic of cataract presence. Conclusions DeepOpacityNet could detect cataracts from CFPs in AREDS2 with performance superior to that of ophthalmologists and generate interpretable results. The code and models are available at https://github.com/ncbi/DeepOpacityNet ( https://doi.org/10.5281/zenodo.10127002 ).

Abstract Deep learning has become a popular tool for computer-aided diagnosis using medical images, sometimes matching or exceeding the performance of clinicians. However, these models can also reflect and amplify human bias, potentially resulting inaccurate missed diagnoses. Despite this concern, the problem of improving model fairness in medical image classification by deep learning has yet to be fully studied. To address this issue, we propose an algorithm that leverages the marginal pairwise equal opportunity to reduce bias in medical image classification. Our evaluations across four tasks using four independent large-scale cohorts demonstrate that our proposed algorithm not only improves fairness in individual and intersectional subgroups but also maintains overall performance. Specifically, the relative change in pairwise fairness difference between our proposed model and the baseline model was reduced by over 35%, while the relative change in AUC value was typically within 1%. By reducing the bias generated by deep learning models, our proposed approach can potentially alleviate concerns about the fairness and reliability of image-based computer-aided diagnosis.

BackgroundAlthough patients have easy access to their electronic health records and laboratory test result data through patient portals, laboratory test results are often confusing and hard to understand. Many patients turn to web-based forums or question-and-answer (Q&A) sites to seek advice from their peers. The quality of answers from social Q&A sites on health-related questions varies significantly, and not all responses are accurate or reliable. Large language models (LLMs) such as ChatGPT have opened a promising avenue for patients to have their questions answered. ObjectiveWe aimed to assess the feasibility of using LLMs to generate relevant, accurate, helpful, and unharmful responses to laboratory test–related questions asked by patients and identify potential issues that can be mitigated using augmentation approaches. MethodsWe collected laboratory test result–related Q&A data from Yahoo! Answers and selected 53 Q&A pairs for this study. Using the LangChain framework and ChatGPT web portal, we generated responses to the 53 questions from 5 LLMs: GPT-4, GPT-3.5, LLaMA 2, MedAlpaca, and ORCA_mini. We assessed the similarity of their answers using standard Q&A similarity-based evaluation metrics, including Recall-Oriented Understudy for Gisting Evaluation, Bilingual Evaluation Understudy, Metric for Evaluation of Translation With Explicit Ordering, and Bidirectional Encoder Representations from Transformers Score. We used an LLM-based evaluator to judge whether a target model had higher quality in terms of relevance, correctness, helpfulness, and safety than the baseline model. We performed a manual evaluation with medical experts for all the responses to 7 selected questions on the same 4 aspects. ResultsRegarding the similarity of the responses from 4 LLMs; the GPT-4 output was used as the reference answer, the responses from GPT-3.5 were the most similar, followed by those from LLaMA 2, ORCA_mini, and MedAlpaca. Human answers from Yahoo data were scored the lowest and, thus, as the least similar to GPT-4–generated answers. The results of the win rate and medical expert evaluation both showed that GPT-4’s responses achieved better scores than all the other LLM responses and human responses on all 4 aspects (relevance, correctness, helpfulness, and safety). LLM responses occasionally also suffered from lack of interpretation in one’s medical context, incorrect statements, and lack of references. ConclusionsBy evaluating LLMs in generating responses to patients’ laboratory test result–related questions, we found that, compared to other 4 LLMs and human answers from a Q&A website, GPT-4’s responses were more accurate, helpful, relevant, and safer. There were cases in which GPT-4 responses were inaccurate and not individualized. We identified a number of ways to improve the quality of LLM responses, including prompt engineering, prompt augmentation, retrieval-augmented generation, and response evaluation.

Summary: Biomedical research yields vast information, much of which is only accessible through the literature. Consequently, literature search is crucial for healthcare and biomedicine. Recent improvements in artificial intelligence (AI) have expanded functionality beyond keywords, but they might be unfamiliar to clinicians and researchers. In response, we present an overview of over 30 literature search tools tailored to common biomedical use cases, aiming at helping readers efficiently fulfill their information needs. We first discuss recent improvements and continued challenges of the widely used PubMed. Then, we describe AI-based literature search tools catering to five specific information needs: 1. Evidence-based medicine. 2. Precision medicine and genomics. 3. Searching by meaning, including questions. 4. Finding related articles with literature recommendation. 5. Discovering hidden associations through literature mining. Finally, we discuss the impacts of recent developments of large language models such as ChatGPT on biomedical information seeking.

In the Jiaoshiba block of the Fuling shale gas field, the employed reserves and recovery factor by primary well pattern are low, no obvious barrier is found in the development layer series, and layered development is difficult. Based on the understanding of the main factors controlling shale gas enrichment and high production, the theory and technology of shale gas three-dimensional development, such as fine description and modeling of shale gas reservoir, optimization of three-dimensional development strategy, highly efficient drilling with dense well pattern, precision fracturing and real-time control, are discussed. Three-dimensional development refers to the application of optimal and fast drilling and volume fracturing technologies, depending upon the sedimentary characteristics, reservoir characteristics and sweet spot distribution of shale gas, to form “artificial gas reservoir” in a multidimensional space, so as to maximize the employed reserves, recovery factor and yield rate of shale gas development. In the research on shale gas three-dimensional development, the geological + engineering sweet spot description is fundamental, the collaborative optimization of natural fractures and artificial fractures is critical, and the improvement of speed and efficiency in drilling and fracturing engineering is the guarantee. Through the implementation of three-dimensional development, the overall recovery factor in the Jiaoshiba block has increased from 12.6% to 23.3%, providing an important support for the continuous and stable production of the Fuling shale gas field.

Abstract Background The Pooled Cohort Equations (PCEs) are race- and sex-specific Cox proportional hazards (PH)-based models used for 10-year atherosclerotic cardiovascular disease (ASCVD) risk prediction with acceptable discrimination. In recent years, neural network models have gained increasing popularity with their success in image recognition and text classification. Various survival neural network models have been proposed by combining survival analysis and neural network architecture to take advantage of the strengths from both. However, the performance of these survival neural network models compared to each other and to PCEs in ASCVD prediction is unknown. Methods In this study, we used 6 cohorts from the Lifetime Risk Pooling Project (with 5 cohorts as training/internal validation and one cohort as external validation) and compared the performance of the PCEs in 10-year ASCVD risk prediction with an all two-way interactions Cox PH model (Cox PH-TWI) and three state-of-the-art neural network survival models including Nnet-survival, Deepsurv, and Cox-nnet. For all the models, we used the same 7 covariates as used in the PCEs. We fitted each of the aforementioned models in white females, white males, black females, and black males, respectively. We evaluated models’ internal and external discrimination power and calibration. Results The training/internal validation sample comprised 23216 individuals. The average age at baseline was 57.8 years old (SD = 9.6); 16% developed ASCVD during average follow-up of 10.50 (SD = 3.02) years. Based on 10 × 10 cross-validation, the method that had the highest C-statistics was Deepsurv (0.7371) for white males, Deepsurv and Cox PH-TWI (0.7972) for white females, PCE (0.6981) for black males, and Deepsurv (0.7886) for black females. In the external validation dataset, Deepsurv (0.7032), Cox-nnet (0.7282), PCE (0.6811), and Deepsurv (0.7316) had the highest C-statistics for white male, white female, black male, and black female population, respectively. Calibration plots showed that in 10 × 10 validation, all models had good calibration in all race and sex groups. In external validation, all models overestimated the risk for 10-year ASCVD. Conclusions We demonstrated the use of the state-of-the-art neural network survival models in ASCVD risk prediction. Neural network survival models had similar if not superior discrimination and calibration compared to PCEs.

Objective: To propose Deep-GA-Net, a 3-dimensional (3D) deep learning network with 3D attention layer, for the detection of geographic atrophy (GA) on spectral domain OCT (SD-OCT) scans, explain its decision making, and compare it with existing methods. Design: Deep learning model development. Participants: Three hundred eleven participants from the Age-Related Eye Disease Study 2 Ancillary SD-OCT Study. Methods: A dataset of 1284 SD-OCT scans from 311 participants was used to develop Deep-GA-Net. Cross-validation was used to evaluate Deep-GA-Net, where each testing set contained no participant from the corresponding training set. En face heatmaps and important regions at the B-scan level were used to visualize the outputs of Deep-GA-Net, and 3 ophthalmologists graded the presence or absence of GA in them to assess the explainability (i.e., understandability and interpretability) of its detections. Main Outcome Measures: Accuracy, area under receiver operating characteristic curve (AUC), area under precision-recall curve (APR). Results: Compared with other networks, Deep-GA-Net achieved the best metrics, with accuracy of 0.93, AUC of 0.94, and APR of 0.91, and received the best gradings of 0.98 and 0.68 on the en face heatmap and B-scan grading tasks, respectively. Conclusions: Deep-GA-Net was able to detect GA accurately from SD-OCT scans. The visualizations of Deep-GA-Net were more explainable, as suggested by 3 ophthalmologists. The code and pretrained models are publicly available at https://github.com/ncbi/Deep-GA-Net. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

During hydraulic fracturing of waxy shale oil reservoirs, the presence of fracturing fluid can influence the phase behavior of the fluid within the reservoir, and heat exchange between the fluids causes wax precipitation that impacts reservoir development. To investigate multiscale fluid phase transition and microscale flow impacted by fracturing fluid injection, this study conducted no-water phase behavior experiments, water injection wax precipitation experiments, and water-condition phase behavior experiments using a nanofluidic chip model. The results show that in the no-water phase experiment, the gasification occurred first in the large cracks, while the matrix throat was the last, and the bubble point pressure difference between the two was 12.1 MPa. The wax precipitation phenomena during fracturing fluid injection can be divided into granular wax in cracks, flake wax in cracks, and wax precipitation in the matrix throat, and the wax mainly accumulated in the microcracks and remained in the form of particles. Compared with the no-water conditions, the large cracks and matrix throat bubble point in the water conditions decreased by 6.1 MPa and 3.5 MPa, respectively, and the presence of the water phase reduced the material occupancy ratio at each pore scale. For the smallest matrix throat, the final gas occupancy ratio under the water conditions decreased from 32% to 24% in the experiment without water. This study provides valuable insight into reservoir fracture modification and guidance for the efficient development of similar reservoirs.

AbstractIn plants, β-glucosidases (BGLUs) are important glycoside hydrolases involved in several biological phenomena, including the response to stresses via the activation of phytohormones and the release of alpha-hydroxy nitriles to protect against stresses. Due to the importance of BGLUs in plant growth and stress response, genome-wide analyses have been conducted in Arabidopsis and rice, while not in tomato (Solanum lycopersicum). In this study, we identified 20 BGLU genes in the tomato that were unevenly distributed on nine chromosomes and divided into five subgroups. There were a variety of plant hormones and stress response cis-elements in the promoter region of the SlBGLU gene, indicating that the BGLU gene may be involved in several aspects of the tomato stress response. SlBGLU gene expression analysis showed that the BGLU gene of tomato was expressed in many tissues, especially in the roots and leaves. Transcriptome data and results of real-time quantitative reverse transcription-polymerase chain reaction showed that most SlBGLU genes could be induced by abscisic acid, chlorothalonil, NaCl and cold (4 °C), especially SlBGLU13 and SlBGLU19, which may be key BGLU genes in response to stress and hormonal stimulation in the tomato. In addition, we constructed a competing endogenous RNA (ceRNA) network, which confers a new direction for studies on the function of SlBGLU genes. These findings not only further clarify the potential function of the BGLU gene family in mediating abiotic stresses in the tomato, but also provide valuable information for the study of functional genomics of the tomato in the future.

Abstract As non-renewable cultural heritage, wall paintings play an important role in society. To reveal the trends in the scientific analysis of mural paintings, 845 relevant research articles published from 2011 to 2021 were collected from the Web of Science database and analyzed. The VOSviewer software was adopted to map the network data of scientific publications, so that relationships among authors, countries, institutions can be displayed, and the co-occurrence of keywords and co-citation can be analyzed. The results revealed close and strong interconnections between the top authors, suggesting a considerable strong research link in this field. The cooperation between research institutions was relatively close. The most productive country of relevant publications was Italy. The leading journals for the scientific analysis of wall paintings were Journal of Raman Spectroscopy and Journal of Cultural Heritage. At present, the hotspots of scientific analysis and research on wall painting are revealing the composition, distribution, origin, and deterioration mechanism of pigments, alongside with evaluating the effects and mechanism of conservation materials and techniques. On the one hand, a possible development direction in this field is introducing more cutting-edge analysis and data processing methods. On the other hand, scientific analysis is increasingly adopted to guide the research and development of mural conservation materials.

To conquer the trade-off dilemma of the nanofiltration membranes for both high permeability for water and high rejection for pollutants has long been highly concerned. In this work, TiO2 and Bi2S3 nanoparticles were evenly dispersed successively into the interlayers of Ti3C2Tx nanosheets, and then filtered onto cellulose acetate support membranes. The as-obtained composite membranes exhibited a maximum water flux of 374.4 L m−2 h−1 bar−1, and a high rejection ratio for RhB, GR, MB and BSA at 97.8–100%, apparently higher than those of the Ti3C2Tx and the related membranes in literature. Such a high separation capacity was found quite stable for various dye-containing aqueous solutions, even in a high salinity environment of 5 wt% NaCl. Interestingly, 25% NaCl salt was also rejected accompanying the almost complete rejection of dye pollutants. After the filtering process, the membrane fouling of fouled Ti3C2Tx/TiO2/Bi2S3 membranes was verified quite lower than that of fouled Ti3C2Tx/TiO2 and Ti3C2Tx membranes. Furthermore, the flux recovery rate of the fouled composite membrane could reach 92.1% after washing under visible light, which could be ascribed to the composition of heterojunction causing the red shift of diffuse reflectance UV–vis spectra and the narrowing of the band gap. This study provides a new idea for the preparation of highly permeable and rejective nanofiltration membranes, and lays a foundation for its high-capacity dye removal applications.

Biochar-immobilized bacteria have been widely used to remove organic pollutants; however, the enhanced effect of biochar-immobilized bacteria on tetrabromobisphenol A (TBBPA) removal has not been fully investigated and the removal mechanism remains unclear. In this study, a bacterial strain with high TBBPA degradation ability, Burkholderia cepacian Y17, was isolated from an e-waste disassembly area, immobilized with biochar, and used for the removal of TBBPA. Comparisons were performed of the factors affecting the immobilization and TBBPA removal efficiency, including the biochar preparation temperature, immobilization temperature, and pH. The highest 7-day TBBPA removal efficiency by immobilized bacteria was observed with the most suitable biochar preparation temperature (BC500) and an immobilization pH and temperature of 7 and 35 °C, respectively. The TBBPA removal efficiency reached 59.37%, which was increased by 30.23% and 15.88% compared to that of free and inactivated immobilized Y17, respectively. The suitable biochar preparation temperature BC500, immobilization temperature of 35 °C, and neutral pH of 7 increased the bacterial population and extracellular polymer concentration, which facilitated bacterial immobilization on biochar and promoted TBBPA removal. In this case, the high immobilized bacteria concentration (4.62 × 108 cfu∙g−1) and protein and polysaccharide contents (28.43 and 16.16 mg·g−1) contributed to the removal of TBBPA by facilitating TBBPA degradation. The main TBBPA degradation processes by BC500-immobilized Y17 involved debromination, β-scission, demethylation, O-methylation, hydroxylation, and hydroxyl oxidation. This study proposes a method for preparing immobilized bacteria for TBBPA removal and enriches the microbial degradation technology for TBBPA.

Measurement(s) chemical entity • textual entity • Annotation Technology Type(s) expert manual annotation • digital curation • machine learning Machine-accessible metadata file describing the reported data: https://doi.org/10.6084/m9.figshare.13486839

Abstract By 2040, age-related macular degeneration (AMD) will affect ~288 million people worldwide. Identifying individuals at high risk of progression to late AMD, the sight-threatening stage, is critical for clinical actions, including medical interventions and timely monitoring. Although deep learning has shown promise in diagnosing/screening AMD using color fundus photographs, it remains difficult to predict individuals’ risks of late AMD accurately. For both tasks, these initial deep learning attempts have remained largely unvalidated in independent cohorts. Here, we demonstrate how deep learning and survival analysis can predict the probability of progression to late AMD using 3298 participants (over 80,000 images) from the Age-Related Eye Disease Studies AREDS and AREDS2, the largest longitudinal clinical trials in AMD. When validated against an independent test data set of 601 participants, our model achieved high prognostic accuracy (5-year C-statistic 86.4 (95% confidence interval 86.2–86.6)) that substantially exceeded that of retinal specialists using two existing clinical standards (81.3 (81.1–81.5) and 82.0 (81.8–82.3), respectively). Interestingly, our approach offers additional strengths over the existing clinical standards in AMD prognosis (e.g., risk ascertainment above 50%) and is likely to be highly generalizable, given the breadth of training data from 82 US retinal specialty clinics. Indeed, during external validation through training on AREDS and testing on AREDS2 as an independent cohort, our model retained substantially higher prognostic accuracy than existing clinical standards. These results highlight the potential of deep learning systems to enhance clinical decision-making in AMD patients.

Abstract As one of the most ubiquitous diagnostic imaging tests in medical practice, chest radiography requires timely reporting of potential findings and diagnosis of diseases in the images. Automated, fast, and reliable detection of diseases based on chest radiography is a critical step in radiology workflow. In this work, we developed and evaluated various deep convolutional neural networks (CNN) for differentiating between normal and abnormal frontal chest radiographs, in order to help alert radiologists and clinicians of potential abnormal findings as a means of work list triaging and reporting prioritization. A CNN-based model achieved an AUC of 0.9824 ± 0.0043 (with an accuracy of 94.64 ± 0.45%, a sensitivity of 96.50 ± 0.36% and a specificity of 92.86 ± 0.48%) for normal versus abnormal chest radiograph classification. The CNN model obtained an AUC of 0.9804 ± 0.0032 (with an accuracy of 94.71 ± 0.32%, a sensitivity of 92.20 ± 0.34% and a specificity of 96.34 ± 0.31%) for normal versus lung opacity classification. Classification performance on the external dataset showed that the CNN model is likely to be highly generalizable, with an AUC of 0.9444 ± 0.0029. The CNN model pre-trained on cohorts of adult patients and fine-tuned on pediatric patients achieved an AUC of 0.9851 ± 0.0046 for normal versus pneumonia classification. Pretraining with natural images demonstrates benefit for a moderate-sized training image set of about 8500 images. The remarkable performance in diagnostic accuracy observed in this study shows that deep CNNs can accurately and effectively differentiate normal and abnormal chest radiographs, thereby providing potential benefits to radiology workflow and patient care.

Abstract Background Capturing sentence semantics plays a vital role in a range of text mining applications. Despite continuous efforts on the development of related datasets and models in the general domain, both datasets and models are limited in biomedical and clinical domains. The BioCreative/OHNLP2018 organizers have made the first attempt to annotate 1068 sentence pairs from clinical notes and have called for a community effort to tackle the Semantic Textual Similarity (BioCreative/OHNLP STS) challenge. Methods We developed models using traditional machine learning and deep learning approaches. For the post challenge, we focused on two models: the Random Forest and the Encoder Network. We applied sentence embeddings pre-trained on PubMed abstracts and MIMIC-III clinical notes and updated the Random Forest and the Encoder Network accordingly. Results The official results demonstrated our best submission was the ensemble of eight models. It achieved a Person correlation coefficient of 0.8328 – the highest performance among 13 submissions from 4 teams. For the post challenge, the performance of both Random Forest and the Encoder Network was improved; in particular, the correlation of the Encoder Network was improved by ~ 13%. During the challenge task, no end-to-end deep learning models had better performance than machine learning models that take manually-crafted features. In contrast, with the sentence embeddings pre-trained on biomedical corpora, the Encoder Network now achieves a correlation of ~ 0.84, which is higher than the original best model. The ensembled model taking the improved versions of the Random Forest and Encoder Network as inputs further increased performance to 0.8528. Conclusions Deep learning models with sentence embeddings pre-trained on biomedical corpora achieve the highest performance on the test set. Through error analysis, we find that end-to-end deep learning models and traditional machine learning models with manually-crafted features complement each other by finding different types of sentences. We suggest a combination of these models can better find similar sentences in practice.

Photocatalytic reduction of carbon dioxide (CO2) into value-added carbon products is a promising strategy for dealing with carbon emission. Selected metal–organic frameworks (MOFs) have been shown promise in photocatalytic CO2 reduction owing to their photochemical activity and structural modularity. Here, we take advantage of displaceable nonstructural ligands on the nodes of a pyrene-based MOF, NU-1000, to graft a molecular catalyst, tetra(4-carboxyphenyl) porphyrin iron (III) chloride (Fe-TCPP). The pyrene linkers in the MOF absorb photons and transfer electrons to the porphyrin catalyst, resulting in a self-contained photocatalyst. The obtained Fe-TCPP@NU-1000 heterogeneous catalysts exhibit activity for CO2 photoreduction under illumination at 390 nm, with a TON for CO of 22 in 2 h. Molecular hydrogen is co-evolved. Leaching of the grafted catalyst from the MOF ultimately limits performance.

We investigated the geometric evolution of mesoscopic mechanical structures in rocks during the rock fragmentation process induced by tunnel boring machine cutters. Numerical models were built using a grain-based discrete element method to accurately represent the mesoscopic structures and macroscopic mechanical behaviours of rocks, and the relationship between the mesoscopic evolution and the macroscopic response of rock was determined. The major results are as follows. First, the crushing and re-compaction of the grains were found to mainly occur in the thin crushed zone immediately beneath the cutter tip. Second, the reduction in the bearing ability of the dense core during cutter indentation was due to the order increment of the contact topological structure at the mesoscopic scale. Third, the area percentages of low- and high-order meso-loops decreased and increased, respectively, during the indentation process, and the volume expansion of the dense core was mainly caused by an increase in the internal pore area of high-order meso-loops that have low internal solid fractions. Fourth, the low-order meso-loops primarily bore and transferred the indentation force. Finally, the distribution contour of the meso-loops was found to be an appropriate and intuitive approach for representing the evolution of cracks on a macroscopic scale.

BackgroundSemantic textual similarity (STS) measures the degree of relatedness between sentence pairs. The Open Health Natural Language Processing (OHNLP) Consortium released an expertly annotated STS data set and called for the National Natural Language Processing Clinical Challenges. This work describes our entry, an ensemble model that leverages a range of deep learning (DL) models. Our team from the National Library of Medicine obtained a Pearson correlation of 0.8967 in an official test set during 2019 National Natural Language Processing Clinical Challenges/Open Health Natural Language Processing shared task and achieved a second rank. ObjectiveAlthough our models strongly correlate with manual annotations, annotator-level correlation was only moderate (weighted Cohen κ=0.60). We are cautious of the potential use of DL models in production systems and argue that it is more critical to evaluate the models in-depth, especially those with extremely high correlations. In this study, we benchmark the effectiveness and efficiency of top-ranked DL models. We quantify their robustness and inference times to validate their usefulness in real-time applications. MethodsWe benchmarked five DL models, which are the top-ranked systems for STS tasks: Convolutional Neural Network, BioSentVec, BioBERT, BlueBERT, and ClinicalBERT. We evaluated a random forest model as an additional baseline. For each model, we repeated the experiment 10 times, using the official training and testing sets. We reported 95% CI of the Wilcoxon rank-sum test on the average Pearson correlation (official evaluation metric) and running time. We further evaluated Spearman correlation, R², and mean squared error as additional measures. ResultsUsing only the official training set, all models obtained highly effective results. BioSentVec and BioBERT achieved the highest average Pearson correlations (0.8497 and 0.8481, respectively). BioSentVec also had the highest results in 3 of 4 effectiveness measures, followed by BioBERT. However, their robustness to sentence pairs of different similarity levels varies significantly. A particular observation is that BERT models made the most errors (a mean squared error of over 2.5) on highly similar sentence pairs. They cannot capture highly similar sentence pairs effectively when they have different negation terms or word orders. In addition, time efficiency is dramatically different from the effectiveness results. On average, the BERT models were approximately 20 times and 50 times slower than the Convolutional Neural Network and BioSentVec models, respectively. This results in challenges for real-time applications. ConclusionsDespite the excitement of further improving Pearson correlations in this data set, our results highlight that evaluations of the effectiveness and efficiency of STS models are critical. In future, we suggest more evaluations on the generalization capability and user-level testing of the models. We call for community efforts to create more biomedical and clinical STS data sets from different perspectives to reflect the multifaceted notion of sentence-relatedness.

Purpose: Manually identifying geographic atrophy (GA) presence and location on OCT volume scans can be challenging and time consuming. This study developed a deep learning model simultaneously (1) to perform automated detection of GA presence or absence from OCT volume scans and (2) to provide interpretability by demonstrating which regions of which B-scans show GA. Design: Med-XAI-Net, an interpretable deep learning model was developed to detect GA presence or absence from OCT volume scans using only volume scan labels, as well as to interpret the most relevant B-scans and B-scan regions. Participants: One thousand two hundred eighty-four OCT volume scans (each containing 100 B-scans) from 311 participants, including 321 volumes with GA and 963 volumes without GA. Methods: Med-XAI-Net simulates the human diagnostic process by using a region-attention module to locate the most relevant region in each B-scan, followed by an image-attention module to select the most relevant B-scans for classifying GA presence or absence in each OCT volume scan. Med-XAI-Net was trained and tested (80% and 20% participants, respectively) using gold standard volume scan labels from human expert graders. Main Outcome Measures: Accuracy, area under the receiver operating characteristic (ROC) curve, F1 score, sensitivity, and specificity. Results: In the detection of GA presence or absence, Med-XAI-Net obtained superior performance (91.5%, 93.5%, 82.3%, 82.8%, and 94.6% on accuracy, area under the ROC curve, F1 score, sensitivity, and specificity, respectively) to that of 2 other state-of-the-art deep learning methods. The performance of ophthalmologists grading only the 5 B-scans selected by Med-XAI-Net as most relevant (95.7%, 95.4%, 91.2%, and 100%, respectively) was almost identical to that of ophthalmologists grading all volume scans (96.0%, 95.7%, 91.8%, and 100%, respectively). Even grading only 1 region in 1 B-scan, the ophthalmologists demonstrated moderately high performance (89.0%, 87.4%, 77.6%, and 100%, respectively). Conclusions: Despite using ground truth labels during training at the volume scan level only, Med-XAI-Net was effective in locating GA in B-scans and selecting relevant B-scans within each volume scan for GA diagnosis. These results illustrate the strengths of Med-XAI-Net in interpreting which regions and B-scans contribute to GA detection in the volume scan.

Data-driven research in biomedical science requires structured, computable data. Increasingly, these data are created with support from automated text mining. Text-mining tools have rapidly matured: although not perfect, they now frequently provide outstanding results. We describe 10 straightforward writing tips-and a web tool, PubReCheck-guiding authors to help address the most common cases that remain difficult for text-mining tools. We anticipate these guides will help authors' work be found more readily and used more widely, ultimately increasing the impact of their work and the overall benefit to both authors and readers. PubReCheck is available at http://www.ncbi.nlm.nih.gov/research/pubrecheck.

A massive number of biological entities, such as genes and mutations, are mentioned in the biomedical literature. The capturing of the semantic relatedness of biological entities is vital to many biological applications, such as protein-protein interaction prediction and literature-based discovery. Concept embeddings-which involve the learning of vector representations of concepts using machine learning models-have been employed to capture the semantics of concepts. To develop concept embeddings, named-entity recognition (NER) tools are first used to identify and normalize concepts from the literature, and then different machine learning models are used to train the embeddings. Despite multiple attempts, existing biomedical concept embeddings generally suffer from suboptimal NER tools, small-scale evaluation, and limited availability. In response, we employed high-performance machine learning-based NER tools for concept recognition and trained our concept embeddings, BioConceptVec, via four different machine learning models on ~30 million PubMed abstracts. BioConceptVec covers over 400,000 biomedical concepts mentioned in the literature and is of the largest among the publicly available biomedical concept embeddings to date. To evaluate the validity and utility of BioConceptVec, we respectively performed two intrinsic evaluations (identifying related concepts based on drug-gene and gene-gene interactions) and two extrinsic evaluations (protein-protein interaction prediction and drug-drug interaction extraction), collectively using over 25 million instances from nine independent datasets (17 million instances from six intrinsic evaluation tasks and 8 million instances from three extrinsic evaluation tasks), which is, by far, the most comprehensive to our best knowledge. The intrinsic evaluation results demonstrate that BioConceptVec consistently has, by a large margin, better performance than existing concept embeddings in identifying similar and related concepts. More importantly, the extrinsic evaluation results demonstrate that using BioConceptVec with advanced deep learning models can significantly improve performance in downstream bioinformatics studies and biomedical text-mining applications. Our BioConceptVec embeddings and benchmarking datasets are publicly available at https://github.com/ncbi-nlp/BioConceptVec.

The exploration and development of highly-efficient Pt-alternative catalysts with acceptable costs are very essential for the large-scale commercial applications of direct methanol fuel cell (DMFC) technology. Herein, we demonstrate a facile and cost-effective self-assembly approach to the bottom-up fabrication of Rh-decorated three-dimensional graphene aerogel (Rh/3D-GA) networks as efficient DMFC anode catalysts. With the distinct textural features including large surface area, 3D porous structure, uniform Rh dispersion, and good electron conductivity, the resultant Rh/3D-GA hybrid architecture shows superior electrocatalytic properties toward methanol oxidation reaction, such as large electrochemically active surface area, low onset potential, high mass activity as well as excellent long-term stability, all of which are more competitive than those of Rh catalysts deposited on conventional carbon black, carbon nanotubes, and two-dimensional graphene supports.

Abstract In order to accurately identify the ancient liquid contained in the plum vase excavated from Jurou Li’s Grave of the Jin Dynasty (1115–1234 CE) in Xi’an, mass spectrometry was applied to determine the amino acid sequences of the residual proteins extracted from the liquid sample. The sequences were searched against a standard protein sequence database. The proteins extracted was identified as glycosyltransferase from Sorghum bicolor, calcium-dependent protein kinase 2 from Wickerhamomyces ciferrii, and cytochrome b-c1 complex subunit Rieske from Nadsonia fulvescens. These findings indicate that the extremely degraded liquid in the plum vase was made from the cereal of sorghum by alcoholic fermentation of Wickerhamomyces ciferrii and Nadsonia fulvescens, providing direct evidence for liquor in the Jin Dynasty.

Background/Aims: The isolation and establishment of female germline stem cells (FGSCs) is controversial because of questions regarding the reliability and stability of the isolation method using antibody targeting mouse vasa homologue (MVH), and the molecular mechanism of FGSCs self-renewal remains unclear. Thus, there needs to be a simple and reliable method for sorting FGSCs to study them. Methods: We applied the differential adhesion method to enrich FGSCs (DA-FGSCs) from mouse ovaries. Through four rounds of purification and 7-9 subsequent passages, DA-FGSC lines were established. In addition, we assessed the role of the phosphoinositide-3 kinase (PI3K)-AKT pathway in regulating FGSC self-renewal. Results: The obtained DA-FGSCs spontaneously differentiated into oocyte-like cells in vitro and formed functional eggs in vivo that were fertilized and produced healthy offspring. AKT was rapidly phosphorylated when the proliferation rate of FGSCs increased after 10 passages, and the addition of a chemical PI3K inhibitor prevented FGSCs self-renewal. Furthermore, over-expression of AKT-induced proliferation and differentiation of FGSCs, c-Myc, Oct-4 and Gdf-9 levels were increased. Conclusions: The differential adhesion method provides a more feasible approach and is an easier procedure to establish FGSC lines than traditional methods. The AKT pathway plays an important role in regulation of the proliferation and maintenance of FGSCs. These findings could help promote stem cell studies and provide a better understanding of causes of ovarian infertility, thereby providing potential treatments for infertility.

Abstract Background MeSH indexing is the task of assigning relevant MeSH terms based on a manual reading of scholarly publications by human indexers. The task is highly important for improving literature retrieval and many other scientific investigations in biomedical research. Unfortunately, given its manual nature, the process of MeSH indexing is both time-consuming (new articles are not immediately indexed until 2 or 3 months later) and costly (approximately ten dollars per article). In response, automatic indexing by computers has been previously proposed and attempted but remains challenging. In order to advance the state of the art in automatic MeSH indexing, a community-wide shared task called BioASQ was recently organized. Methods We propose MeSH Now, an integrated approach that first uses multiple strategies to generate a combined list of candidate MeSH terms for a target article. Through a novel learning-to-rank framework, MeSH Now then ranks the list of candidate terms based on their relevance to the target article. Finally, MeSH Now selects the highest-ranked MeSH terms via a post-processing module. Results We assessed MeSH Now on two separate benchmarking datasets using traditional precision, recall and F1-score metrics. In both evaluations, MeSH Now consistently achieved over 0.60 in F-score, ranging from 0.610 to 0.612. Furthermore, additional experiments show that MeSH Now can be optimized by parallel computing in order to process MEDLINE documents on a large scale. Conclusions We conclude that MeSH Now is a robust approach with state-of-the-art performance for automatic MeSH indexing and that MeSH Now is capable of processing PubMed scale documents within a reasonable time frame. Availability: http://www.ncbi.nlm.nih.gov/CBBresearch/Lu/Demo/MeSHNow/ .

Manually curating biomedical knowledge from publications is necessary to build a knowledge based service that provides highly precise and organized information to users. The process of retrieving relevant publications for curation, which is also known as document triage, is usually carried out by querying and reading articles in PubMed. However, this query-based method often obtains unsatisfactory precision and recall on the retrieved results, and it is difficult to manually generate optimal queries. To address this, we propose a machine-learning assisted triage method. We collect previously curated publications from two databases UniProtKB/Swiss-Prot and the NHGRI-EBI GWAS Catalog, and used them as a gold-standard dataset for training deep learning models based on convolutional neural networks. We then use the trained models to classify and rank new publications for curation. For evaluation, we apply our method to the real-world manual curation process of UniProtKB/Swiss-Prot and the GWAS Catalog. We demonstrate that our machine-assisted triage method outperforms the current query-based triage methods, improves efficiency, and enriches curated content. Our method achieves a precision 1.81 and 2.99 times higher than that obtained by the current query-based triage methods of UniProtKB/Swiss-Prot and the GWAS Catalog, respectively, without compromising recall. In fact, our method retrieves many additional relevant publications that the query-based method of UniProtKB/Swiss-Prot could not find. As these results show, our machine learning-based method can make the triage process more efficient and is being implemented in production so that human curators can focus on more challenging tasks to improve the quality of knowledge bases.

PubMed is a free search engine for biomedical literature accessed by millions of users from around the world each day. With the rapid growth of biomedical literature-about two articles are added every minute on average-finding and retrieving the most relevant papers for a given query is increasingly challenging. We present Best Match, a new relevance search algorithm for PubMed that leverages the intelligence of our users and cutting-edge machine-learning technology as an alternative to the traditional date sort order. The Best Match algorithm is trained with past user searches with dozens of relevance-ranking signals (factors), the most important being the past usage of an article, publication date, relevance score, and type of article. This new algorithm demonstrates state-of-the-art retrieval performance in benchmarking experiments as well as an improved user experience in real-world testing (over 20% increase in user click-through rate). Since its deployment in June 2017, we have observed a significant increase (60%) in PubMed searches with relevance sort order: it now assists millions of PubMed searches each week. In this work, we hope to increase the awareness and transparency of this new relevance sort option for PubMed users, enabling them to retrieve information more effectively.

Staff from the National Center for Biotechnology Information in the US describe recent improvements to the PubMed search engine and outline plans for the future, including a new experimental site called PubMed Labs.

The practice of precision medicine will ultimately require databases of genes and mutations for healthcare providers to reference in order to understand the clinical implications of each patient's genetic makeup. Although the highest quality databases require manual curation, text mining tools can facilitate the curation process, increasing accuracy, coverage, and productivity. However, to date there are no available text mining tools that offer high-accuracy performance for extracting such triplets from biomedical literature. In this paper we propose a high-performance machine learning approach to automate the extraction of disease-gene-variant triplets from biomedical literature. Our approach is unique because we identify the genes and protein products associated with each mutation from not just the local text content, but from a global context as well (from the Internet and from all literature in PubMed). Our approach also incorporates protein sequence validation and disease association using a novel text-mining-based machine learning approach. We extract disease-gene-variant triplets from all abstracts in PubMed related to a set of ten important diseases (breast cancer, prostate cancer, pancreatic cancer, lung cancer, acute myeloid leukemia, Alzheimer's disease, hemochromatosis, age-related macular degeneration (AMD), diabetes mellitus, and cystic fibrosis). We then evaluate our approach in two ways: (1) a direct comparison with the state of the art using benchmark datasets; (2) a validation study comparing the results of our approach with entries in a popular human-curated database (UniProt) for each of the previously mentioned diseases. In the benchmark comparison, our full approach achieves a 28% improvement in F1-measure (from 0.62 to 0.79) over the state-of-the-art results. For the validation study with UniProt Knowledgebase (KB), we present a thorough analysis of the results and errors. Across all diseases, our approach returned 272 triplets (disease-gene-variant) that overlapped with entries in UniProt and 5,384 triplets without overlap in UniProt. Analysis of the overlapping triplets and of a stratified sample of the non-overlapping triplets revealed accuracies of 93% and 80% for the respective categories (cumulative accuracy, 77%). We conclude that our process represents an important and broadly applicable improvement to the state of the art for curation of disease-gene-variant relationships.

Female germline stem cells (FGSCs) or oogonial stem cells (OSCs) have the capacity to generate newborn oocytes and thus open a new door to fight ovarian aging and female infertility. However, the production and identification of OSCs are difficult for investigators. Rare amount of these cells in the ovary results in the failure of the acquisition of OSCs. Furthermore, the oocyte formation by OSCs in vivo was usually confirmed using tissue sections by immunofluorescence or immunohistochemistry in previous studies. STO or MEF feeder cells are derived from mouse, not human. In our study, we modified the protocol. The cells were digested from ovaries and cultured for 2-3 days and then were purified by magnetic-activated cell sorting (MACS). The ovaries and fetus of mice injected with EGFP-positive OSCs were prepared and put on the slides to directly visualize oocyte and progeny formation under microscope. Additionally, the human umbilical cord mesenchymal stem cells (hUC-MSCs) were also used as feeder cells to support the proliferation of OSCs. The results showed that all the modified procedures can significantly improve and facilitate the generation and characterization of OSCs, and hUC-MSCs as feeder will be useful for isolation and proliferation of human OSCs avoiding contamination from mouse.

The discovery of new uses of existing drugs offers the possibility to reduce time and risk because the approved drugs have passed several phases along the drug development pipeline. In this study, we present a computational method for novel drug use prediction based on the idea that similar drugs are indicated for similar diseases. When computing drug pairwise similarities, we considered both chemical structure and drug target similarities. In validation, our new drug use predictions were found to be significantly enriched in both the biomedical literature and clinical trials. These results indicate that our method is able to successfully integrate both biomedical scientific data and literature for drug discovery.

Early menopause and infertility often occur in female cancer patients after chemotherapy (CTx). For these patients, oocyte/embryo cryopreservation or ovarian tissue cryopreservation is the current modality for fertility preservation. However, the above methods are limited in the long-term protection of ovarian function, especially for fertility preservation (very few females with cancer have achieved pregnancy with cryopreserved ovarian tissue or eggs until now). In addition, the above methods are subject to their scope (females with no husband or prepubertal females with no mature oocytes). Thus, many females who suffer from cancers would not adopt the above methods pre- and post-CTx due to their uncertainty, safety and cost-effectiveness. Therefore, millions of women have achieved long-term survival after thorough CTx treatment and have desired to rescue their ovarian function and fertility with economic, durable and reliable methods. Recently, some studies showed that mice with infertility caused by CTx can produce normal offspring through intraovarian injection of exogenous female germline stem cells (FGSCs). Though exogenous FGSC can be derived from mice without immune rejection in the same strain, it is difficult to obtain human female germline stem cells (hFGSCs), and immune rejection could occur between different individuals. In this study, infertility in mice was caused by CTx, and the ability of FGSCs to restore ovarian function or even produce offspring was assessed. We had successfully isolated and purified the FGSCs from adult female mice two weeks after CTx. After infection with GFP-carrying virus, the FGSCs were transplanted into ovaries of mice with infertility caused by CTx. Finally, ovarian function was restored and the recipients produced offspring long-term. These findings showed that mice with CTx possessed FGSCs, restoring ovarian function and avoiding immune rejection from exogenous germline stem cells.