HTINet2: herb-target prediction via knowledge graph embedding and residual-like graph neural network.

Target identification is one of the crucial tasks in drug research and development, as it aids in uncovering the action mechanism of herbs/drugs and discovering new therapeutic targets. Although multiple algorithms of herb target prediction have been proposed, due to the incompleteness of clinical knowledge and the limitation of unsupervised models, accurate identification for herb targets still faces huge challenges of data and models. To address this, we proposed a deep learning-based target prediction framework termed HTINet2, which designed three key modules, namely, traditional Chinese medicine (TCM) and clinical knowledge graph embedding, residual graph representation learning, and supervised target prediction. In the first module, we constructed a large-scale knowledge graph that covers the TCM properties and clinical treatment knowledge of herbs, and designed a component of deep knowledge embedding to learn the deep knowledge embedding of herbs and targets. In the remaining two modules, we designed a residual-like graph convolution network to capture the deep interactions among herbs and targets, and a Bayesian personalized ranking loss to conduct supervised training and target prediction. Finally, we designed comprehensive experiments, of which comparison with baselines indicated the excellent performance of HTINet2 (HR@10 increased by 122.7% and NDCG@10 by 35.7%), ablation experiments illustrated the positive effect of our designed modules of HTINet2, and case study demonstrated the reliability of the predicted targets of Artemisia annua and Coptis chinensis based on the knowledge base, literature, and molecular docking.

KDGene: knowledge graph completion for disease gene prediction using interactional tensor decomposition.

The accurate identification of disease-associated genes is crucial for understanding the molecular mechanisms underlying various diseases. Most current methods focus on constructing biological networks and utilizing machine learning, particularly deep learning, to identify disease genes. However, these methods overlook complex relations among entities in biological knowledge graphs. Such information has been successfully applied in other areas of life science research, demonstrating their effectiveness. Knowledge graph embedding methods can learn the semantic information of different relations within the knowledge graphs. Nonetheless, the performance of existing representation learning techniques, when applied to domain-specific biological data, remains suboptimal. To solve these problems, we construct a biological knowledge graph centered on diseases and genes, and develop an end-to-end knowledge graph completion framework for disease gene prediction using interactional tensor decomposition named KDGene. KDGene incorporates an interaction module that bridges entity and relation embeddings within tensor decomposition, aiming to improve the representation of semantically similar concepts in specific domains and enhance the ability to accurately predict disease genes. Experimental results show that KDGene significantly outperforms state-of-the-art algorithms, whether existing disease gene prediction methods or knowledge graph embedding methods for general domains. Moreover, the comprehensive biological analysis of the predicted results further validates KDGene's capability to accurately identify new candidate genes. This work proposes a scalable knowledge graph completion framework to identify disease candidate genes, from which the results are promising to provide valuable references for further wet experiments. Data and source codes are available at https://github.com/2020MEAI/KDGene.

DRONet: effectiveness-driven drug repositioning framework using network embedding and ranking learning.

As one of the most vital methods in drug development, drug repositioning emphasizes further analysis and research of approved drugs based on the existing large amount of clinical and experimental data to identify new indications of drugs. However, the existing drug repositioning methods didn't achieve enough prediction performance, and these methods do not consider the effectiveness information of drugs, which make it difficult to obtain reliable and valuable results. In this study, we proposed a drug repositioning framework termed DRONet, which make full use of effectiveness comparative relationships (ECR) among drugs as prior information by combining network embedding and ranking learning. We utilized network embedding methods to learn the deep features of drugs from a heterogeneous drug-disease network, and constructed a high-quality drug-indication data set including effectiveness-based drug contrast relationships. The embedding features and ECR of drugs are combined effectively through a designed ranking learning model to prioritize candidate drugs. Comprehensive experiments show that DRONet has higher prediction accuracy (improving 87.4% on Hit@1 and 37.9% on mean reciprocal rank) than state of the art. The case analysis also demonstrates high reliability of predicted results, which has potential to guide clinical drug development.

TMNP: a transcriptome-based multi-scale network pharmacology platform for herbal medicine.

One of the most difficult problems that hinder the development and application of herbal medicine is how to illuminate the global effects of herbs on the human body. Currently, the chemo-centric network pharmacology methodology regards herbs as a mixture of chemical ingredients and constructs the 'herb-compound-target-disease' connections based on bioinformatics methods, to explore the pharmacological effects of herbal medicine. However, this approach is severely affected by the complexity of the herbal composition. Alternatively, gene-expression profiles induced by herbal treatment reflect the overall biological effects of herbs and are suitable for studying the global effects of herbal medicine. Here, we develop an online transcriptome-based multi-scale network pharmacology platform (TMNP) for exploring the global effects of herbal medicine. Firstly, we build specific functional gene signatures for different biological scales from molecular to higher tissue levels. Then, specific algorithms are designed to measure the correlations of transcriptional profiles and types of gene signatures. Finally, TMNP uses pharmacotranscriptomics of herbal medicine as input and builds associations between herbs and different biological scales to explore the multi-scale effects of herb medicine. We applied TMNP to a single herb Astragalus membranaceus and Xuesaitong injection to demonstrate the power to reveal the multi-scale effects of herbal medicine. TMNP integrating herbal medicine and multiple biological scales into the same framework, will greatly extend the conventional network pharmacology model centering on the chemical components, and provide a window for systematically observing the complex interactions between herbal medicine and the human body. TMNP is available at http://www.bcxnfz.top/TMNP.

A network-based machine-learning framework to identify both functional modules and disease genes.

Disease gene identification is a critical step towards uncovering the molecular mechanisms of diseases and systematically investigating complex disease phenotypes. Despite considerable efforts to develop powerful computing methods, candidate gene identification remains a severe challenge owing to the connectivity of an incomplete interactome network, which hampers the discovery of true novel candidate genes. We developed a network-based machine-learning framework to identify both functional modules and disease candidate genes. In this framework, we designed a semi-supervised non-negative matrix factorization model to obtain the functional modules related to the diseases and genes. Of note, we proposed a disease gene-prioritizing method called MapGene that integrates the correlations from both functional modules and network closeness. Our framework identified a set of functional modules with highly functional homogeneity and close gene interactions. Experiments on a large-scale benchmark dataset showed that MapGene performs significantly better than the state-of-the-art algorithms. Further analysis demonstrates MapGene can effectively relieve the impact of the incompleteness of interactome networks and obtain highly reliable rankings of candidate genes. In addition, disease cases on Parkinson's disease and diabetes mellitus confirmed the generalization of MapGene for novel candidate gene identification. This work proposed, for the first time, an integrated computing framework to predict both functional modules and disease candidate genes. The methodology and results support that our framework has the potential to help discover underlying functional modules and reliable candidate genes in human disease.

SymMap: an integrative database of traditional Chinese medicine enhanced by symptom mapping.

Recently, the pharmaceutical industry has heavily emphasized phenotypic drug discovery (PDD), which relies primarily on knowledge about phenotype changes associated with diseases. Traditional Chinese medicine (TCM) provides a massive amount of information on natural products and the clinical symptoms they are used to treat, which are the observable disease phenotypes that are crucial for clinical diagnosis and treatment. Curating knowledge of TCM symptoms and their relationships to herbs and diseases will provide both candidate leads and screening directions for evidence-based PDD programs. Therefore, we present SymMap, an integrative database of traditional Chinese medicine enhanced by symptom mapping. We manually curated 1717 TCM symptoms and related them to 499 herbs and 961 symptoms used in modern medicine based on a committee of 17 leading experts practicing TCM. Next, we collected 5235 diseases associated with these symptoms, 19 595 herbal constituents (ingredients) and 4302 target genes, and built a large heterogeneous network containing all of these components. Thus, SymMap integrates TCM with modern medicine in common aspects at both the phenotypic and molecular levels. Furthermore, we inferred all pairwise relationships among SymMap components using statistical tests to give pharmaceutical scientists the ability to rank and filter promising results to guide drug discovery. The SymMap database can be accessed at http://www.symmap.org/ and https://www.bioinfo.org/symmap.

Longitudinal clinical trajectory analysis of individuals before and after diagnosis of Type 2 Diabetes Mellitus (T2DM) indicates that vascular problems start early.

INTRODUCTION: Type 2 diabetes mellitus (T2DM) frequently associates with increasing multi-morbidity/treatment complexity. Some headway has been made to identify genetic and non-genetic risk factors for T2DM. However, longitudinal clinical histories of individuals both before and after diagnosis of T2DM are likely to provide additional insight into both diabetes aetiology/further complex trajectory of multi-morbidity. METHODS: This study utilised diabetes patients/controls enrolled in the DARE (Diabetes Alliance for Research in England) study where pre- and post-T2DM diagnosis longitudinal data was available for trajectory analysis. Longitudinal data of 281 individuals (T2DM n = 237 vs matched non-T2DM controls n = 44) were extracted, checked for errors and logical inconsistencies and then subjected to Trajectory Analysis over a period of up to 70 years based on calculations of the proportions of most prominent clinical conditions for each year. RESULTS: For individuals who eventually had a diagnosis of T2DM made, a number of clinical phenotypes were seen to increase consistently in the years leading up to diagnosis of T2DM. Of these documented phenotypes, the most striking were diagnosed hypertension (more than in the control group) and asthma. This trajectory over time was much less dramatic in the matched control group. Immediately prior to T2DM diagnosis, a greater indication of ischaemic heart disease proportions was observed. Post-T2DM diagnosis, the proportions of T2DM patients exhibiting hypertension and infection continued to climb rapidly before plateauing. Ischaemic heart disease continued to increase in this group as well as retinopathy, impaired renal function and heart failure. CONCLUSION: These observations provide an intriguing and novel insight into the onset and natural progression of T2DM. They suggest an early phase of potentially related disease activity well before any clinical diagnosis of diabetes is made. Further studies on a larger cohort of DARE patients are underway to explore the utility of establishing predictive risk scores.

Identification of herbal categories active in pain disorder subtypes by machine learning help reveal novel molecular mechanisms of algesia.

Chronic pain is highly prevalent and poorly controlled, of which the accurate underlying mechanisms need be further elucidated. Herbal drugs have been widely used for controlling various pain disorders. The systematic integration of pain herbal data resources might be promising to help investigate the molecular mechanisms of pain phenotypes. Here, we integrated large-scale bibliographic literatures and well-established data sources to obtain high-quality pain relevant herbal data (i.e. 426 pain related herbs with their targets). We used machine learning method to identify three distinct herb categories with their specific indications of symptoms, targets and enriched pathways, which were characterized by the efficacy of treatment to the chronic cough related neuropathic pain, the reproduction and autoimmune related pain, and the cancer pain, respectively. We further detected the novel pathophysiological mechanisms of the pain subtypes by network medicine approach to evaluate the interactions between herb targets and the pain disease modules. This work increased the understanding of the underlying molecular mechanisms of pain subtypes that herbal drugs are participating and with the ultimate aim of developing novel personalized drugs for pain disorders.

Integrated network analysis of symptom clusters across disease conditions.

Identifying the symptom clusters (two or more related symptoms) with shared underlying molecular mechanisms has been a vital analysis task to promote the symptom science and precision health. Related studies have applied the clustering algorithms (e.g. k-means, latent class model) to detect the symptom clusters mostly from various kinds of clinical data. In addition, they focused on identifying the symptom clusters (SCs) for a specific disease, which also mainly concerned with the clinical regularities for symptom management. Here, we utilized a network-based clustering algorithm (i.e., BigCLAM) to obtain 208 typical SCs across disease conditions on a large-scale symptom network derived from integrated high-quality disease-symptom associations. Furthermore, we evaluated the underlying shared molecular mechanisms for SCs, i.e., shared genes, protein-protein interaction (PPI) and gene functional annotations using integrated networks and similarity measures. We found that the symptoms in the same SCs tend to share a higher degree of genes, PPIs and have higher functional homogeneities. In addition, we found that most SCs have related symptoms with shared underlying molecular mechanisms (e.g. enriched pathways) across different disease conditions. Our work demonstrated that the integrated network analysis method could be used for identifying robust SCs and investigate the molecular mechanisms of these SCs, which would be valuable for symptom science and precision health.

Genetics of rheumatoid arthritis contributes to biology and drug discovery.

A major challenge in human genetics is to devise a systematic strategy to integrate disease-associated variants with diverse genomic and biological data sets to provide insight into disease pathogenesis and guide drug discovery for complex traits such as rheumatoid arthritis (RA). Here we performed a genome-wide association study meta-analysis in a total of >100,000 subjects of European and Asian ancestries (29,880 RA cases and 73,758 controls), by evaluating ∼10 million single-nucleotide polymorphisms. We discovered 42 novel RA risk loci at a genome-wide level of significance, bringing the total to 101 (refs 2 - 4). We devised an in silico pipeline using established bioinformatics methods based on functional annotation, cis-acting expression quantitative trait loci and pathway analyses--as well as novel methods based on genetic overlap with human primary immunodeficiency, haematological cancer somatic mutations and knockout mouse phenotypes--to identify 98 biological candidate genes at these 101 risk loci. We demonstrate that these genes are the targets of approved therapies for RA, and further suggest that drugs approved for other indications may be repurposed for the treatment of RA. Together, this comprehensive genetic study sheds light on fundamental genes, pathways and cell types that contribute to RA pathogenesis, and provides empirical evidence that the genetics of RA can provide important information for drug discovery.

An analysis of human microbe-disease associations.

The microbiota living in the human body has critical impacts on our health and disease, but a systems understanding of its relationships with disease remains limited. Here, we use a large-scale text mining-based manually curated microbe-disease association data set to construct a microbe-based human disease network and investigate the relationships between microbes and disease genes, symptoms, chemical fragments and drugs. We reveal that microbe-based disease loops are significantly coherent. Microbe-based disease connections have strong overlaps with those constructed by disease genes, symptoms, chemical fragments and drugs. Moreover, we confirm that the microbe-based disease analysis is able to predict novel connections and mechanisms for disease, microbes, genes and drugs. The presented network, methods and findings can be a resource helpful for addressing some issues in medicine, for example, the discovery of bench knowledge and bedside clinical solutions for disease mechanism understanding, diagnosis and therapy.

Finite element analysis of the effect of power arm locations on tooth movement in extraction space closure with miniscrew anchorage in customized lingual orthodontic treatment.

INTRODUCTION: More patients are choosing customized orthodontic appliances because of their excellent esthetics. It is essential that clinicians understand the biomechanics of the tooth movement tendency in customized lingual orthodontics. This study aimed to evaluate the tooth movement tendency during space closure in maxillary anterior teeth with the use of miniscrew anchorage in customized lingual orthodontics with various power arm locations. METHODS: Three-dimensional finite element models of the maxilla were created with miniscrews and power arms; the positions were varied to change the force directions. A retraction force (1.5 N) was applied from the top of the miniscrews to the selected points on the power arm, and the initial displacements of the reference nodes of the maxillary teeth were analyzed. RESULTS: After applying force in different directions, power arms located at the distal side of the canines led to larger initial lingual crown tipping and occlusal crown extrusion of the maxillary incisors compared with power arms located at the midpoint between the lateral incisors and canines, and caused a decreasing trend of the intercanine width. CONCLUSIONS: In customized lingual orthodontic treatment, power arms located at the distal side of the canines are unfavorable for anterior teeth torque control and intercanine width control. Power arms located at the midpoint between the lateral incisors and canines can get better torque control, but still cannot achieve excepted torque without extra torque control methods, no matter whether its force application point is higher than, lower than, or equal to the level of the top of the miniscrews.

Risk factors associated with 31-day unplanned readmission in 50,912 discharged patients after stroke in China.

BACKGROUND: Unplanned readmission within 31 days of discharge after stroke is a useful indicator for monitoring quality of hospital care. We evaluated the risk factors associated with 31-day unplanned readmission of stroke patients in China. METHODS: We identified 50,912 patients from 375 hospitals in 29 provinces, municipalities or autonomous districts across China who experienced an unplanned readmission after stroke between 2015 and 2016, and extracted data from the inpatients' cover sheet data from the Medical Record Monitoring Database. Patients were grouped into readmission within 31 days or beyond for analysis. Chi-squared test was used to analyze demographic information, health system and clinical process-related factors according to the data type. Multilevel logistic modeling was used to examine the effects of patient (level 1) and hospital (level 2) characteristics on an unplanned readmission ≤31 days. RESULTS: Among 50,912 patients, 14,664 (28.8%) were readmitted within 31 days after discharge. The commonest cause of readmissions were recurrent stroke (34.8%), hypertension (22.94%), cardio/cerebrovascular disease (13.26%) and diabetes/diabetic complications (7.34%). Higher risks of unplanned readmissions were associated with diabetes (OR = 1.089, P = 0.001), use of clinical pathways (OR = 1.174, P < 0.001), and being discharged without doctor's advice (OR = 1.485, P < 0.001). Lower risks were associated with basic medical insurances (OR ranging from 0.225 to 0.716, P < 0.001) and commercial medical insurance (OR = 0.636, P = 0.021), compared to self-paying for medical services. And patients aged 50 years old and above (OR ranging from 0.650 to 0.985, P < 0.05), with haemorrhagic stroke (OR = 0.467, P < 0.001), with length of stay more than 7 days in hospital (OR ranging from 0.082 to 0.566, P < 0.001), also had lower risks. CONCLUSIONS: Age, type of stroke, medical insurance status, type of discharge, use of clinical pathways, length of hospital stay and comorbidities were the most influential factors for readmission within 31 days.

Heterogeneous network propagation for herb target identification.

BACKGROUND: Identifying targets of herbs is a primary step for investigating pharmacological mechanisms of herbal drugs in Traditional Chinese medicine (TCM). Experimental targets identification of herbs is a difficult and time-consuming work. Computational method for identifying herb targets is an efficient approach. However, how to make full use of heterogeneous network data about herbs and targets to improve the performance of herb targets prediction is still a dilemma. METHODS: In our study, a random walk algorithm on the heterogeneous herb-target network (named heNetRW) has been proposed to identify protein targets of herbs. By building a heterogeneous herb-target network involving herbs, targets and their interactions and simulating random walk algorithm on the network, the candidate targets of the given herb can be predicted. RESULTS: The experimental results on large-scale dataset showed that heNetRW had higher performance of targets prediction than PRINCE (improved F1-score by 0.08 and Hit@1 by 21.34% in one validation setting, and improved F1-score by 0.54 and Hit@1 by 69.08% in the other validation setting). Furthermore, we evaluated novel candidate targets of two herbs (rhizoma coptidis and turmeric), which showed our approach could generate potential targets that are valuable for further experimental investigations. CONCLUSIONS: Compared with PRINCE algorithm, heNetRW algorithm can fuse more known information (such as, known herb-target associations and pathway-based similarities of protein pairs) to improve prediction performance. Experimental results also indicated heNetRW had higher performance than PRINCE. The prediction results not only can be used to guide the selection of candidate targets of herbs, but also help to reveal the molecule mechanisms of herbal drugs.

Human genetics in rheumatoid arthritis guides a high-throughput drug screen of the CD40 signaling pathway.

Although genetic and non-genetic studies in mouse and human implicate the CD40 pathway in rheumatoid arthritis (RA), there are no approved drugs that inhibit CD40 signaling for clinical care in RA or any other disease. Here, we sought to understand the biological consequences of a CD40 risk variant in RA discovered by a previous genome-wide association study (GWAS) and to perform a high-throughput drug screen for modulators of CD40 signaling based on human genetic findings. First, we fine-map the CD40 risk locus in 7,222 seropositive RA patients and 15,870 controls, together with deep sequencing of CD40 coding exons in 500 RA cases and 650 controls, to identify a single SNP that explains the entire signal of association (rs4810485, P = 1.4×10(-9)). Second, we demonstrate that subjects homozygous for the RA risk allele have ∼33% more CD40 on the surface of primary human CD19+ B lymphocytes than subjects homozygous for the non-risk allele (P = 10(-9)), a finding corroborated by expression quantitative trait loci (eQTL) analysis in peripheral blood mononuclear cells from 1,469 healthy control individuals. Third, we use retroviral shRNA infection to perturb the amount of CD40 on the surface of a human B lymphocyte cell line (BL2) and observe a direct correlation between amount of CD40 protein and phosphorylation of RelA (p65), a subunit of the NF-κB transcription factor. Finally, we develop a high-throughput NF-κB luciferase reporter assay in BL2 cells activated with trimerized CD40 ligand (tCD40L) and conduct an HTS of 1,982 chemical compounds and FDA-approved drugs. After a series of counter-screens and testing in primary human CD19+ B cells, we identify 2 novel chemical inhibitors not previously implicated in inflammation or CD40-mediated NF-κB signaling. Our study demonstrates proof-of-concept that human genetics can be used to guide the development of phenotype-based, high-throughput small-molecule screens to identify potential novel therapies in complex traits such as RA.

[Combination rules of Chinese herbal prescriptions for treating unstable angina based on complex network].

OBJECTIVE: To explore combination rules of Chinese herbal prescriptions from effective cases for treatment of unstable angina (UA). METHODS: Prescription data from 156 UA patients effectively treated at Cardiovascular Diseases Centre of Xiyuan Hospital were analyzed using complex network method. RESULTS: According to multi-scale analysis of backbone network and pointwise mutual information analysis, core prescriptions from the 156 UA patients were presented as follows: Rhizoma Ligustici wallichii, Radix Paeoniae rubra, Radix Codonopsis, Rhizoma Pinelliae, poria, and Angelica sinensis. Meanwhile, core couplet medicines for these patients covered Rhizoma Ligustici wallichii and Radix paeoniaerubra, Angelica sinensis and Rhizoma Ligustici wallichii, Radix Codonopsis and Rhizoma Ligustici wallichii, Rhizoma Ligustici wallichii and Rhizoma Pinelliae, Rhizoma Atractylodis Macrocephalae and poriacocos, Bulbus Alli Macrostemi and Rhizoma Pinelliae. Among different primary symptoms, there was slightly difference in core prescriptions. CONCLUSION: The core prescriptions for the treatment of UA include blood-activating drug, phlem-resolving drugs. As an exploration of combination rules of Chinese herbal prescriptions in treating UA based on complex network, it can be used as a reference for further researches.

Intelligent quality control of traditional chinese medical tongue diagnosis images based on deep learning.

BACKGROUND: Computer-aided tongue and face diagnosis technology can make Traditional Chinese Medicine (TCM) more standardized, objective and quantified. However, many tongue images collected by the instrument may not meet the standard in clinical applications, which affects the subsequent quantitative analysis. The common tongue diagnosis instrument cannot determine whether the patient has fully extended the tongue or collected the face. OBJECTIVE: This paper proposes an image quality control algorithm based on deep learning to verify the eligibility of TCM tongue diagnosis images. METHODS: We firstly gathered enough images and categorized them into five states. Secondly, we preprocessed the training images. Thirdly, we built a ResNet34 model and trained it by the transfer learning method. Finally, we input the test images into the trained model and automatically filter out unqualified images and point out the reasons. RESULTS: Experimental results show that the model's quality control accuracy rate of the test dataset is as high as 97.06%. Our methods have the strong discriminative power of the learned representation. Compared with previous studies, it can guarantee subsequent tongue image processing. CONCLUSIONS: Our methods can guarantee the subsequent quantitative analysis of tongue shape, tongue state, tongue spirit, and facial complexion.

Soul: An OCTA dataset based on Human Machine Collaborative Annotation Framework.

Branch retinal vein occlusion (BRVO) is the most prevalent retinal vascular disease that constitutes a threat to vision due to increased venous pressure caused by venous effluent in the space, leading to impaired visual function. Optical Coherence Tomography Angiography (OCTA) is an innovative non-invasive technique that offers high-resolution three-dimensional structures of retinal blood vessels. Most publicly available datasets are collected from single visits with different patients, encompassing various eye diseases for distinct tasks and areas. Moreover, due to the intricate nature of eye structure, professional labeling not only relies on the expertise of doctors but also demands considerable time and effort. Therefore, we have developed a BRVO-focused dataset named Soul (Source of ocular vascular) and propose a human machine collaborative annotation framework (HMCAF) using scrambled retinal blood vessels data. Soul is categorized into 6 subsets based on injection frequency and follow-up duration. The dataset comprises original images, corresponding blood vessel labels, and clinical text information sheets which can be effectively utilized when combined with machine learning.

Deep representation learning from electronic medical records identifies distinct symptom based subtypes and progression patterns for COVID-19 prognosis.

OBJECTIVE: Symptoms are significant kind of phenotypes for managing and controlling of the burst of acute infectious diseases, such as COVID-19. Although patterns of symptom clusters and time series have been considered the high potential prediction factors for the prognosis of patients, the elaborated subtypes and their progression patterns based on symptom phenotypes related to the prognosis of COVID-19 patients still need be detected. This study aims to investigate patient subtypes and their progression patterns with distinct features of outcome and prognosis. METHODS: This study included a total of 14,139 longitudinal electronic medical records (EMRs) obtained from four hospitals in Hubei Province, China, involving 2,683 individuals in the early stage of COVID-19 pandemic. A deep representation learning model was developed to help acquire the symptom profiles of patients. K-means clustering algorithm is used to divide them into distinct subtypes. Subsequently, symptom progression patterns were identified by considering the subtypes associated with patients upon admission and discharge. Furthermore, we used Fisher's test to identify significant clinical entities for each subtype. RESULTS: Three distinct patient subtypes exhibiting specific symptoms and prognosis have been identified. Particularly, Subtype 0 includes 44.2% of the whole and is characterized by poor appetite, fatigue and sleep disorders; Subtype 1 includes 25.6% cases and is characterized by confusion, cough with bloody sputum, encopresis and urinary incontinence; Subtype 2 includes 30.2% cases and is characterized by dry cough and rhinorrhea. These three subtypes demonstrate significant disparities in prognosis, with the mortality rates of 4.72%, 8.59%, and 0.25% respectively. Furthermore, symptom cluster progression patterns showed that patients with Subtype 0 who manifest dark yellow urine, chest pain, etc. in the admission stage exhibit an elevated risk of transforming into the more severe subtypes with poor outcome, whereas those presenting with nausea and vomiting tend to incline towards entering the milder subtype. CONCLUSION: This study has proposed a clinical meaningful approach by utilizing the deep representation learning and real-world EMR data containing symptom phenotypes to identify the COVID-19 subtypes and their progression patterns. The results would be potentially useful to help improve the precise stratification and management of acute infectious diseases.

PresRecST: a novel herbal prescription recommendation algorithm for real-world patients with integration of syndrome differentiation and treatment planning.

OBJECTIVES: Herbal prescription recommendation (HPR) is a hot topic and challenging issue in field of clinical decision support of traditional Chinese medicine (TCM). However, almost all previous HPR methods have not adhered to the clinical principles of syndrome differentiation and treatment planning of TCM, which has resulted in suboptimal performance and difficulties in application to real-world clinical scenarios. MATERIALS AND METHODS: We emphasize the synergy among diagnosis and treatment procedure in real-world TCM clinical settings to propose the PresRecST model, which effectively combines the key components of symptom collection, syndrome differentiation, treatment method determination, and herb recommendation. This model integrates a self-curated TCM knowledge graph to learn the high-quality representations of TCM biomedical entities and performs 3 stages of clinical predictions to meet the principle of systematic sequential procedure of TCM decision making. RESULTS: To address the limitations of previous datasets, we constructed the TCM-Lung dataset, which is suitable for the simultaneous training of the syndrome differentiation, treatment method determination, and herb recommendation. Overall experimental results on 2 datasets demonstrate that the proposed PresRecST outperforms the state-of-the-art algorithm by significant improvements (eg, improvements of P@5 by 4.70%, P@10 by 5.37%, P@20 by 3.08% compared with the best baseline). DISCUSSION: The workflow of PresRecST effectively integrates the embedding vectors of the knowledge graph for progressive recommendation tasks, and it closely aligns with the actual diagnostic and treatment procedures followed by TCM doctors. A series of ablation experiments and case study show the availability and interpretability of PresRecST, indicating the proposed PresRecST can be beneficial for assisting the diagnosis and treatment in real-world TCM clinical settings. CONCLUSION: Our technology can be applied in a progressive recommendation scenario, providing recommendations for related items in a progressive manner, which can assist in providing more reliable diagnoses and herbal therapies for TCM clinical task.