Adaptive evolution of the enigmatic Takakia now facing climate change in Tibet.

The most extreme environments are the most vulnerable to transformation under a rapidly changing climate. These ecosystems harbor some of the most specialized species, which will likely suffer the highest extinction rates. We document the steepest temperature increase (2010-2021) on record at altitudes of above 4,000 m, triggering a decline of the relictual and highly adapted moss Takakia lepidozioides. Its de-novo-sequenced genome with 27,467 protein-coding genes includes distinct adaptations to abiotic stresses and comprises the largest number of fast-evolving genes under positive selection. The uplift of the study site in the last 65 million years has resulted in life-threatening UV-B radiation and drastically reduced temperatures, and we detected several of the molecular adaptations of Takakia to these environmental changes. Surprisingly, specific morphological features likely occurred earlier than 165 mya in much warmer environments. Following nearly 400 million years of evolution and resilience, this species is now facing extinction.

Reversible acetylation of HDAC8 regulates cell cycle.

HDAC8, a member of class I HDACs, plays a pivotal role in cell cycle regulation by deacetylating the cohesin subunit SMC3. While cyclins and CDKs are well-established cell cycle regulators, our knowledge of other regulators remains limited. Here we reveal the acetylation of K202 in HDAC8 as a key cell cycle regulator responsive to stress. K202 acetylation in HDAC8, primarily catalyzed by Tip60, restricts HDAC8 activity, leading to increased SMC3 acetylation and cell cycle arrest. Furthermore, cells expressing the mutant form of HDAC8 mimicking K202 acetylation display significant alterations in gene expression, potentially linked to changes in 3D genome structure, including enhanced chromatid loop interactions. K202 acetylation impairs cell cycle progression by disrupting the expression of cell cycle-related genes and sister chromatid cohesion, resulting in G2/M phase arrest. These findings indicate the reversible acetylation of HDAC8 as a cell cycle regulator, expanding our understanding of stress-responsive cell cycle dynamics.

INTEDE 2.0: the metabolic roadmap of drugs.

The metabolic roadmap of drugs (MRD) is a comprehensive atlas for understanding the stepwise and sequential metabolism of certain drug in living organisms. It plays a vital role in lead optimization, personalized medication, and ADMET research. The MRD consists of three main components: (i) the sequential catalyses of drug and its metabolites by different drug-metabolizing enzymes (DMEs), (ii) a comprehensive collection of metabolic reactions along the entire MRD and (iii) a systematic description on efficacy & toxicity for all metabolites of a studied drug. However, there is no database available for describing the comprehensive metabolic roadmaps of drugs. Therefore, in this study, a major update of INTEDE was conducted, which provided the stepwise & sequential metabolic roadmaps for a total of 4701 drugs, and a total of 22 165 metabolic reactions containing 1088 DMEs and 18 882 drug metabolites. Additionally, the INTEDE 2.0 labeled the pharmacological properties (pharmacological activity or toxicity) of metabolites and provided their structural information. Furthermore, 3717 drug metabolism relationships were supplemented (from 7338 to 11 055). All in all, INTEDE 2.0 is highly expected to attract broad interests from related research community and serve as an essential supplement to existing pharmaceutical/biological/chemical databases. INTEDE 2.0 can now be accessible freely without any login requirement at: http://idrblab.org/intede/.

The Effect of Technology-Based Home Cardiac Rehabilitation on Risk Factor Modifications in Coronary Heart Disease Patients. A Systematic Review and Meta-Analysis.

Background: The delivery channels and approaches related to cardiac rehabilitation (CR), such as eHealth, mHealth, and telehealth, are evolving. Several studies have identified their effects on patients with coronary heart disease, although no studies have focused on all the approaches collectively. Methods: Randomized controlled trials have investigated lipid profiles, through systolic blood pressure (SBP), diastolic blood pressure (DBP), and body mass index (BMI). Stata software was used for analysis, while Egger's linear regression test and Begg's funnel plot were also applied. Results: Technology-based home CR revealed significantly lower total cholesterol (TC) levels (standardized mean difference (SMD) = -0.19; 95% confidence interval [CI]: [-0.27, -0.11]); triglyceride (TG) levels (SMD = -0.26; 95% CI: [-0.35, 0.17]); low-density lipoprotein (LDL) levels (SMD = -0.18; 95% CI: [ -0.25, -0.11]); SBP (SMD = -0.26; 95% CI: [-0.33, -0.19]); DBP (SMD = -0.24; 95% CI: [-0.32, -0.16]); BMI (SMD = -0.12; 95% CI: [-0.18, -0.05]), and improved high-density lipoprotein (HDL) levels (SMD = 0.22; 95% CI: [0.14, 0.31]). Conclusions: Technology-based home CR can be used to lower TC, TG, and LDL levels, alongside the BMI, SBP, and DBP indexes, while also raising HDL levels; thus, its use should be widely promoted.

Liver-specific deletion of microRNA-34a alleviates ductular reaction and liver fibrosis during experimental cholestasis.

Primary sclerosing cholangitis (PSC) is a chronic liver disease characterized by inflammatory responses and fibrotic scar formation leading to cholestasis. Ductular reaction and liver fibrosis are typical liver changes seen in human PSC and cholestasis patients. The current study aimed to clarify the role of liver-specific microRNA-34a in the cholestasis-associated ductular reaction and liver fibrosis. We demonstrated that miR-34a expression was significantly increased in human PSC livers along with the enhanced ductular reaction, cellular senescence, and liver fibrosis. A liver-specific miR-34a knockout mouse was established by crossing floxed miR-34a mice with albumin-promoter-driven Cre mice. Bile duct ligation (BDL) induced liver injury characterized by necrosis, fibrosis, and immune cell infiltration. In contrast, liver-specific miR-34a knockout in BDL mice resulted in decreased biliary ductular pathology associated with the reduced cholangiocyte senescence and fibrotic responses. The miR-34a-mediated ductular reactions may be functioning through Sirt-1-mediated senescence and fibrosis. The hepatocyte-derived conditioned medium promoted LPS-induced fibrotic responses and senescence in cholangiocytes, and miR-34a inhibitor suppressed these effects, further supporting the involvement of paracrine regulation. In conclusion, we demonstrated that liver-specific miR-34a plays an important role in ductular reaction and fibrotic responses in a BDL mouse model of cholestatic liver disease.

Discovery of Potent Selective HDAC6 Inhibitors with 5-Phenyl-1H-indole Fragment: Virtual Screening, Rational Design, and Biological Evaluation.

Among the HDACs family, histone deacetylase 6 (HDAC6) has attracted extensive attention due to its unique structure and biological functions. Numerous studies have shown that compared with broad-spectrum HDACs inhibitors, selective HDAC6 inhibitors exert ideal efficacy in tumor treatment with insignificant toxic and side effects, demonstrating promising clinical application prospect. Herein, we carried out rational drug design by integrating a deep learning model, molecular docking, and molecular dynamics simulation technology to construct a virtual screening process. The designed derivatives with 5-phenyl-1H-indole fragment as Cap showed desirable cytotoxicity to the various tumor cell lines, all of which were within 15 µM (ranging from 0.35 to 14.87 µM), among which compound 5i had the best antiproliferative activities against HL-60 (IC50 = 0.35 ± 0.07 µM) and arrested HL-60 cells in the G0/G1 phase. In addition, 5i exhibited better isotype selective inhibitory activities due to the potent potency against HDAC6 (IC50 = 5.16 ± 0.25 nM) and the reduced inhibitory activities against HDAC1 (selective index ≈ 124), which was further verified by immunoblotting results. Moreover, the representative binding conformation of 5i on HDAC6 was revealed and the key residues contributing 5i's binding were also identified via decomposition free-energy analysis. The discovery of lead compound 5i also indicates that virtual screening is still a beneficial tool in drug discovery and can provide more molecular skeletons with research potential for drug design, which is worthy of widespread application.

VARIDT 2.0: structural variability of drug transporter.

The structural variability data of drug transporter (DT) are key for research on precision medicine and rational drug use. However, these valuable data are not sufficiently covered by the available databases. In this study, a major update of VARIDT (a database previously constructed to provide DTs' variability data) was thus described. First, the experimentally resolved structures of all DTs reported in the original VARIDT were discovered from PubMed and Protein Data Bank. Second, the structural variability data of each DT were collected by literature review, which included: (a) mutation-induced spatial variations in folded state, (b) difference among DT structures of human and model organisms, (c) outward/inward-facing DT conformations and (d) xenobiotics-driven alterations in the 3D complexes. Third, for those DTs without experimentally resolved structural variabilities, homology modeling was further applied as well-established protocol to enrich such valuable data. As a result, 145 mutation-induced spatial variations of 42 DTs, 1622 inter-species structures originating from 292 DTs, 118 outward/inward-facing conformations belonging to 59 DTs, and 822 xenobiotics-regulated structures in complex with 57 DTs were updated to VARIDT (https://idrblab.org/varidt/ and http://varidt.idrblab.net/). All in all, the newly collected structural variabilities will be indispensable for explaining drug sensitivity/selectivity, bridging preclinical research with clinical trial, revealing the mechanism underlying drug-drug interaction, and so on.

Development and validation a prognostic model based on natural killer T cells marker genes for predicting prognosis and characterizing immune status in glioblastoma through integrated analysis of single-cell and bulk RNA sequencing.

BACKGROUND: Glioblastoma (GBM) is an aggressive and unstoppable malignancy. Natural killer T (NKT) cells, characterized by specific markers, play pivotal roles in many tumor-associated pathophysiological processes. Therefore, investigating the functions and complex interactions of NKT cells is great interest for exploring GBM. METHODS: We acquired a single-cell RNA-sequencing (scRNA-seq) dataset of GBM from Gene Expression Omnibus (GEO) database. The weighted correlation network analysis (WGCNA) was employed to further screen genes subpopulations. Subsequently, we integrated the GBM cohorts from The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) databases to describe different subtypes by consensus clustering and developed a prognostic model by least absolute selection and shrinkage operator (LASSO) and multivariate Cox regression analysis. We further investigated differences in survival rates and clinical characteristics among different risk groups. Furthermore, a nomogram was developed by combining riskscore with the clinical characteristics. We investigated the abundance of immune cells in the tumor microenvironment (TME) by CIBERSORT and single sample gene set enrichment analysis (ssGSEA) algorithms. Immunotherapy efficacy assessment was done with the assistance of Tumor Immune Dysfunction and Exclusion (TIDE) and The Cancer Immunome Atlas (TCIA) databases. Real-time quantitative polymerase chain reaction (RT-qPCR) experiments and immunohistochemical profiles of tissues were utilized to validate model genes. RESULTS: We identified 945 NKT cells marker genes from scRNA-seq data. Through further screening, 107 genes were accurately identified, of which 15 were significantly correlated with prognosis. We distinguished GBM samples into two distinct subtypes and successfully developed a robust prognostic prediction model. Survival analysis indicated that high expression of NKT cell marker genes was significantly associated with poor prognosis in GBM patients. Riskscore can be used as an independent prognostic factor. The nomogram was demonstrated remarkable utility in aiding clinical decision making. Tumor immune microenvironment analysis revealed significant differences of immune infiltration characteristics between different risk groups. In addition, the expression levels of immune checkpoint-associated genes were consistently elevated in the high-risk group, suggesting more prominent immune escape but also a stronger response to immune checkpoint inhibitors. CONCLUSIONS: By integrating scRNA-seq and bulk RNA-seq data analysis, we successfully developed a prognostic prediction model that incorporates two pivotal NKT cells marker genes, namely, CD44 and TNFSF14. This model has exhibited outstanding performance in assessing the prognosis of GBM patients. Furthermore, we conducted a preliminary investigation into the immune microenvironment across various risk groups that contributes to uncover promising immunotherapeutic targets specific to GBM.

Deep drug-target binding affinity prediction with multiple attention blocks.

Drug-target interaction (DTI) prediction has drawn increasing interest due to its substantial position in the drug discovery process. Many studies have introduced computational models to treat DTI prediction as a regression task, which directly predict the binding affinity of drug-target pairs. However, existing studies (i) ignore the essential correlations between atoms when encoding drug compounds and (ii) model the interaction of drug-target pairs simply by concatenation. Based on those observations, in this study, we propose an end-to-end model with multiple attention blocks to predict the binding affinity scores of drug-target pairs. Our proposed model offers the abilities to (i) encode the correlations between atoms by a relation-aware self-attention block and (ii) model the interaction of drug representations and target representations by the multi-head attention block. Experimental results of DTI prediction on two benchmark datasets show our approach outperforms existing methods, which are benefit from the correlation information encoded by the relation-aware self-attention block and the interaction information extracted by the multi-head attention block. Moreover, we conduct the experiments on the effects of max relative position length and find out the best max relative position length value $k \in \{3, 5\}$. Furthermore, we apply our model to predict the binding affinity of Corona Virus Disease 2019 (COVID-19)-related genome sequences and $3137$ FDA-approved drugs.

Comprehensive analysis identifies cuproptosis-related gene DLAT as a potential prognostic and immunological biomarker in pancreatic adenocarcinoma.

BACKGROUND: Cuproptosis is a regulated cell death form associated with tumor progression, clinical outcomes, and immune response. However, the role of cuproptosis in pancreatic adenocarcinoma (PAAD) remains unclear. This study aims to investigate the implications of cuproptosis-related genes (CRGs) in PAAD by integrated bioinformatic methods and clinical validation. METHODS: Gene expression data and clinical information were downloaded from UCSC Xena platform. We analyzed the expression, mutation, methylation, and correlations of CRGs in PAAD. Then, based on the expression profiles of CRGs, patients were divided into 3 groups by consensus clustering algorithm. Dihydrolipoamide acetyltransferase (DLAT) was chosen for further exploration, including prognostic analysis, co-expression analysis, functional enrichment analysis, and immune landscape analysis. The DLAT-based risk model was established by Cox and LASSO regression analysis in the training cohort, and then verified in the validation cohort. Quantitative reverse transcriptase polymerase chain reaction (RT-qPCR) and immunohistochemistry (IHC) assays were performed to examine the expression levels of DLAT in vitro and in vivo, respectively. RESULTS: Most CRGs were highly expressed in PAAD. Among these genes, increased DLAT could serve as an independent risk factor for survival. Co-expression network and functional enrichment analysis indicated that DLAT was engaged in multiple tumor-related pathways. Moreover, DLAT expression was positively correlated with diverse immunological characteristics, such as immune cell infiltration, cancer-immunity cycle, immunotherapy-predicted pathways, and inhibitory immune checkpoints. Submap analysis demonstrated that DLAT-high patients were more responsive to immunotherapeutic agents. Notably, the DLAT-based risk score model possessed high accuracy in predicting prognosis. Finally, the upregulated expression of DLAT was verified by RT-qPCR and IHC assays. CONCLUSIONS: We developed a DLAT-based model to predict patients' clinical outcomes and demonstrated that DLAT was a promising prognostic and immunological biomarker in PAAD, thereby providing a new possibility for tumor therapy.

Endothelial dysfunction in pathological processes of chronic liver disease during aging.

Aging is associated with gradual changes in liver structure and physiological/pathological functions in hepatic cells including hepatocytes, cholangiocytes, Kupffer cells, hepatic stellate cells (HSCs), and liver sinusoidal endothelial cells (LSECs). LSECs are specialized hepatic endothelial cells that regulate liver homeostasis. These cells actively impact the hepatic microenvironment as they have fenestrations and a thin morphology to allow substance exchange between circulating blood and the liver tissue. As aging occurs, LSECs have a reduction in both the number and size of fenestrations, which is referred to as pseudocapillarization. This along with the aging of the liver leads to increased oxidative stress, decreased availability of nitric oxide, decreased hepatic blood flow, and increased inflammatory cytokines in LSECs. Vascular aging can also lead to hepatic hypoxia, HSC activation, and liver fibrosis. In this review, we described the basic structure of LSECs, and the effect of LSECs on hepatic inflammation and fibrosis during aging process. We briefly summarized the changes of hepatic microcirculation during liver inflammation, the effect of aging on the clearance function of LSECs, the interactions between LSECs and immunity, hepatocytes or other hepatic nonparenchymal cells, and the therapeutic intervention of liver diseases by targeting LSECs and vascular system. Since LSECs play an important role in the development of liver fibrosis and the changes of LSEC phenotype occur in the early stage of liver fibrosis, the study of LSECs in the fibrotic liver is valuable for the detection of early liver fibrosis and the early intervention of fibrotic response.

Exploring Health Literacy Categories in Patients With Heart Failure: A Latent Class Analysis.

BACKGROUND: Although a growing number of studies have demonstrated that patients' health literacy is associated with health outcomes, the exact relationship between them is not clear. AIMS AND OBJECTIVES: The aim of this study was to explore latent classes of health literacy in patients with heart failure and analyze the differences among different groups. DESIGN AND METHODS: This is a cross-sectional survey. Patients diagnosed with heart failure were selected from 3 tertiary hospitals in Tianjin, China, from March 2019 to November 2019. We measured patients' health literacy using the Health Literacy Scale for Chronic Patients. Latent class analysis was carried out based on the patients' Health Literacy Scale for Chronic Patients scores. Multinomial logistic regression was used to identify the predictive indicators of the latent classes. RESULTS: The health literacy of patients with heart failure was divided into 3 different latent classes, named "high health literacy group," "low literacy high dependence group," and "moderate literacy high willingness group." There were statistically significant differences in gender, age, smoking history, marital status, education level, household income level, and quality of life among different health literacy classes. Low education level and household income level predicted poor health literacy. CONCLUSION: There were 3 latent classes for the health literacy of patients with heart failure. Different health literacy classes exhibited their own distinctive characteristics. Patients in the "moderate literacy high willingness group" had the worst quality of life. Understanding the specific types of health literacy in patients with heart failure facilitates targeted nursing interventions to improve their quality of life.

Dioscin induces osteosarcoma cell apoptosis by upregulating ROS-mediated P38 MAPK signaling.

Osteosarcoma is the most common primary malignant bone tumor in children and adolescents. Many patients with osteosarcoma readily develop resistance to chemotherapy and have an extremely dismal prognosis. Dioscin, a saponin, is known to exhibit potent anticancer activities and induce cellular death of a variety of cancer types. However, the inhibitory effect of dioscin on osteosarcoma cells and its underlying mechanisms have not been fully elucidated. We investigated the responses of human U2-OS and MG63 osteosarcoma cells to dioscin with regard to proliferation, apoptosis, migration, and invasion, and studied the effect of dioscin on MAPK-related proteins by western blot analysis assays. Dioscin inhibited osteosarcoma cell proliferation, migration, and invasion. Moreover, it induced osteosarcoma cell apoptosis via reactive oxygen species (ROS)-dependent apoptotic signaling. N-acetylcysteine, a reactive oxygen species inhibitor, suppressed dioscin-induced apoptosis, indicating that ROS play an essential role in dioscin-induced apoptosis. Western blot analysis assays showed that p38 MAPK was upregulated after dioscin treatment, and that dioscin induced apoptosis by upregulating ROS-mediated p38 MAPK signaling. Our study suggests that dioscin possesses antitumor activities against human osteosarcoma cells, inhibits osteosarcoma cell proliferation, migration and invasion, and induces osteosarcoma cell apoptosis through upregulating ROS-mediated p38 MAPK signaling. This study may provide a new therapeutic strategy and potential clinical applications for the treatment of osteosarcoma.

Do biological activities of selective histone deacetylase 6 (HDAC6) inhibitors rely on the modification of cap group?

Nowadays, significant progress has been made in the development of selective histone deacetylase 6 (HDAC6) inhibitors, exerting great potential in the treatment of various malignant tumors and neurodegenerative diseases. Previously, selective inhibitory activities of HDAC inhibitors were generally considered sensitive to the interactions between the Cap group and the binding site of HDAC6, and a large number of selective HDAC6 inhibitors have been designed and synthesized based on the strategy. However, some inhibitors without Cap group could also exhibit excellent potency and selective inhibition towards HDAC6, and in this study, BRD9757 and compound 8, as capless selective HDAC6 inhibitors, were selected as molecular probes to explore the difference of their binding interactions in HDAC1&6. Through the analysis of binding-free energies and conformational rearrangements after 1 µs molecular dynamics simulation, it could be learned that although the residues in the binding site remained highly consistent, the binding mechanisms of BRD9757 and compound 8 in HDAC1&6 were different, which will provide valuable hints for the discovery of novel selective HDAC6 inhibitors.

In situ siloxane passivation of colloidal lead halide perovskite via hot injection for light-emitting diodes.

All-inorganic cesium lead halide perovskite (CsPbX3; X = Cl, Br) nanocrystals (NCs) are synthesized via a modified hot injection method using 3-mercaptopropyltrimethoxysilane (MPTMS), together with oleic acid and oleylamine, for in situ passivation of the surface defects. The surface chemistry, revealed by Fourier transform infrared spectroscopy (FTIR) and x-ray photoelectron spectroscopy (XPS) techniques, shows an absence of Si-O-Si network and C-O groups on these in situ passivated CsPbX3 NCs, denoted as InMP-CsPbX3, which is in strong contrast to the counterpart NCs obtained via a postsynthesis exchange strategy. The x-ray diffraction (XRD) pattern indicates a lattice structure significantly strained from the cubic structure. The synthesis of these InMP-CsPbX3 NCs is highly reproducible, and the colloids are stable in nonpolar solvents. The emission wavelength of CsPb(Cl/Br)3 mixed halide perovskite NCs is tuned from 405 nm to 508 nm by reducing the nominal Cl/Br ratio, while the photoluminescence quantum yield (PLQY) is greatly enhanced over the whole spectral range. More importantly, the InMP-treatment is among the few strategies that are promising for electroluminescence in light-emitting diodes.

Metal-enhanced fluorescence of graphene oxide sheets.

Graphene oxide (GO) is an excellent chemical tunable optical platform for imaging and sensing. The photoluminescence (PL) quantum yield of GO is relatively low, which limited the application of the intrinsic and tunable fluorescence from GO. Here, we report the first case of metal-enhanced fluorescence (MEF) of GO. A significant enhancement (~10-fold) in fluorescence intensity is observed from GO on the Ag substrate as compared to that on the glass. FL, Raman, and SEM images are used to investigate the MEF behavior and are coincident with each other. The influence of the metal particle size of Ag substrate is investigated. The fluorescence is also found to be responsive when adding different metal ions into GO solution. GO contacting directly with metal substrate exhibits strong MEF without quenching, which makes it possible to use GO sheets for three-dimension optical imaging and sensing.

Design, Synthesis, and biological evaluation of HDAC6 inhibitors based on Cap modification strategy.

The abnormal biological functions of HDAC6 were closely related to the occurrence and development of various tumors, making HDAC6 gradually become promising therapeutic target for cancer treatment and inspiring researchers to explore and develop selective HDAC inhibitors. In this study, based on the classical pharmacophore model of HDAC inhibitors, 20 compounds were designed and synthesized by modifying the Cap group, and the biological activities of the target compounds were assessed through anti-proliferation and enzyme inhibition experiments. The title compounds exhibited varying degrees of inhibitory activities against the selected tumor cell lines, especially the compounds 9m, 9q, and 12c, which were further evaluated at the enzymatic level. The enzyme inhibition assay showed that compound 12c exerted broad-spectrum enzyme inhibitory activities and compounds 9m and 9q were more inclined to inhibit HDAC6, exhibiting certain selective inhibitory activities among the representative subtypes. Moreover, the binding modes of compounds 9q and 12c in HDAC1&6 were further explored via computational approaches to elucidate the molecular mechanisms underlying selective inhibitory activities, providing valuable hints for the discovery of novel HDAC6 inhibitors.

Improving patient self-description in Chinese online consultation using contextual prompts.

BACKGROUND: Online health care consultation has been widely adopted to supplement traditional face-to-face patient-doctor interactions. Patients benefit from this new modality of consultation because it allows for time flexibility by eliminating the distance barrier. However, unlike the traditional face-to-face approach, the success of online consultation heavily relies on the accuracy of patient-reported conditions and symptoms. The asynchronous interaction pattern further requires clear and effective patient self-description to avoid lengthy conversation, facilitating timely support for patients. METHOD: Inspired by the observation that doctors talk to patients with the goal of eliciting information to reduce uncertainty about patients' conditions, we proposed and evaluated a machine learning-based computational model towards this goal. Key components of the model include (1) how a doctor diagnoses (predicts) a disease given natural language description of a patient's conditions, (2) how to measure if the patient's description is incomplete or more information is needed from the patient; and (3) given the patient's current description, what further information is needed to help a doctor reach a diagnosis decision. This model makes it possible for an online consultation system to immediately prompt a patient to provide more information if it senses that the current description is insufficient. RESULTS: We evaluated the proposed method by using classification-based metrics (accuracy, macro-averaged F-score, area under the receiver operating characteristics curve, and Matthews correlation coefficient) and an uncertainty-based metric (entropy) on three Chinese online consultation corpora. When there was one consultation round, our method delivered better disease prediction performance than the baseline method (No Prompts) and two heuristic methods (Uncertainty-based Prompts and Certainty-based Prompts). CONCLUSION: The disease prediction performance correlated with uncertainty of patients' self-described symptoms and conditions. However, heuristic solutions ignored the context to decrease large amounts of uncertainty, which did not improve the prediction performance. By elaborate design, a machine-learning algorithm can learn the inner connection between a patient's self-description and the specific information doctors need from doctor-patient conversations to provide prompts, which can enrich the information in patient self-description for a better performance in disease prediction, thereby achieving online consultation with fewer rounds of doctor-patient conversation.

When BERT meets Bilbo: a learning curve analysis of pretrained language model on disease classification.

BACKGROUND: Natural language processing (NLP) tasks in the health domain often deal with limited amount of labeled data due to high annotation costs and naturally rare observations. To compensate for the lack of training data, health NLP researchers often have to leverage knowledge and resources external to a task at hand. Recently, pretrained large-scale language models such as the Bidirectional Encoder Representations from Transformers (BERT) have been proven to be a powerful way of learning rich linguistic knowledge from massive unlabeled text and transferring that knowledge to downstream tasks. However, previous downstream tasks often used training data at such a large scale that is unlikely to obtain in the health domain. In this work, we aim to study whether BERT can still benefit downstream tasks when training data are relatively small in the context of health NLP. METHOD: We conducted a learning curve analysis to study the behavior of BERT and baseline models as training data size increases. We observed the classification performance of these models on two disease diagnosis data sets, where some diseases are naturally rare and have very limited observations (fewer than 2 out of 10,000). The baselines included commonly used text classification models such as sparse and dense bag-of-words models, long short-term memory networks, and their variants that leveraged external knowledge. To obtain learning curves, we incremented the amount of training examples per disease from small to large, and measured the classification performance in macro-averaged [Formula: see text] score. RESULTS: On the task of classifying all diseases, the learning curves of BERT were consistently above all baselines, significantly outperforming them across the spectrum of training data sizes. But under extreme situations where only one or two training documents per disease were available, BERT was outperformed by linear classifiers with carefully engineered bag-of-words features. CONCLUSION: As long as the amount of training documents is not extremely few, fine-tuning a pretrained BERT model is a highly effective approach to health NLP tasks like disease classification. However, in extreme cases where each class has only one or two training documents and no more will be available, simple linear models using bag-of-words features shall be considered.

Functional specificity, diversity, and redundancy of Arabidopsis JAZ family repressors in jasmonate and COI1-regulated growth, development, and defense.

In response to jasmonates (JAs), the JA receptor Coronatine Insensitive 1 (COI1) recruits JA-zinc-finger inflorescence meristem (ZIM)-domain (JAZ) family repressors for destruction to regulate plant growth, development, and defense. As Arabidopsis encodes 13 JAZ repressors, their functional specificity, diversity, and redundancy in JA/COI1-mediated responses remain unclear. We generated a broad range of jaz mutants based on their phylogenetic relationship to investigate their roles in JA responses. The group I JAZ6 may play an inhibitory role in resistance to Botrytis cinerea, group II (JAZ10)/III (JAZ11/12) in JA-regulated root growth inhibition and susceptibility to Pseudomonas syringae pv tomato DC3000, and group IV JAZ3/4/9 in flowering time delay and defense against insects. JAZs exhibit high redundancy in apical hook curvature. The undecuple jaz1/2/3/4/5/6/7/9/10/11/12 (jaz1-7,9-12) mutations enhance JA responses and suppress the phenotypes of coi1-1 in flowering time, rosette growth, and defense. The JA hypersensitivity of jaz1-7,9-12 in root growth, hook curvature, and leaf yellowing is blocked by coi1-1. jaz1-7,9-12 does not influence the stamen phenotypes of wild-type and coi1-1. jaz1-7,9-12 affects JA-regulated transcriptional profile and recovers a fraction of that in coi1-1. This study contributes to elucidating the specificity, diversity, and redundancy of JAZ members in JA/COI1-regulated growth, development, and defense responses.