Due to the high dimensionality and complexity of the data, the analysis of spatial transcriptome data has been a challenging problem. Meanwhile, cluster analysis is the core issue of the analysis of spatial transcriptome data. In this article, a deep learning approach is proposed based on graph attention networks for clustering analysis of spatial transcriptome data. Our method first enhances the spatial transcriptome data, then uses graph attention networks to extract features from nodes, and finally uses the Leiden algorithm for clustering analysis. Compared with the traditional non-spatial and spatial clustering methods, our method has better performance in data analysis through the clustering evaluation index. The experimental results show that the proposed method can effectively cluster spatial transcriptome data and identify different spatial domains, which provides a new tool for studying spatial transcriptome data.
ObjectiveTo analyze the current situation and international research focuses on the study of medical device risk management. MethodsTo retrieve medical device risk management literature information cited from 2002 to 2011 in PubMed such as high-frequency MeSH; analyze current situation and research focuses of medical device risk management by using bibliometrics, bibliographic item co-occurrence matrix builder (BICOMB), and graphical clustering toolkit (gCluto) for quantitative analysis, high-frequency MeSH term papers cluster visualization analysis. ResultsA total of 7 073 published studies were retrieved, basically suggesting a gradually increasing trend of the number of published papers. The top 3 numbers of first authors' papers referred to three countries: the United States, Britain and Germany, while China ranked twelfth. The top 3 numbers of journal articles referred to the United States, Britain and Holland, while China ranked twenty-second. Twenty journals published more than 50 papers, and all these journals were clinical journals. Thirty-three authors published no less than 5 papers, with the maximum of 18 articles. Totally, there were 124 highfrequency MeSHs. The high-frequency MeSHs were classified into 6 categories by using double cluster analysis: kinds 0 to 4 included risk report, risk analysis, risk assessment and methodology of heart valve prosthesis, coronary stents, peripheral vascular stents, implantable defibrillators and other life support device, surgical repair surgical flaps and minimal invasion surgical device such as laparoscopy; kind 5 focused on safety management, risk control, organization and implementation and other related research based on prevention and control of medical device adverse reaction, medical errors, occupation exposure, and equipment failure. ConclusionThe analysis on international literature on medical device risk management basically shows a gradually increasing trend; most studies published in the clinical medicine journals; research focus on risk assessment, safety management and quality improvement in the application such as angioplasty, artificial prosthesis replacement, plastic surgery, minimally invasive surgery and critical care medicine, and radiology diagnosis and treatment; implantable, life-supported invasive and radiological devices as the main research subject; and characteristics include closely combination between medical device risk management and the application of safe and effective, quality improvement systems for clinical diagnosis and treatment.
ObjectiveTo explore the current status and tendency of the application of CT or MRI in the pancreatic pseudocyst using bibiometric analysis for relative documents, and provide reference information for the future research of radiology. MethodsBibliographies from research literatures of CT or MRI application in the pancreatic pseudocyst from January 1, 2003 to September 20, 2014 in PubMed database were downloaded.The publication years, journals, the first authors, and the frequency of subject headings and subheadings were extracted from them by Bicomb 2.0 software.The subject headings and subheadings appeared more than five times were intercepted as high frequency ones, then created the high frequency subject headings and subheadings co-occurrence matrix.SPSS 22.0 statistical software was applied for clustering analysis with this matrix, then got the major hotspots. ResultsA total of 342 literatures were screened out.The research of CT or MRI application in the pancreatic pseudocyst increased slowly year by year in recent 10 years, then slowly decreased after 2008 year.The related literatures were published in the 164 journals, in which 16 journals (115 literatures were published) were core area distribution according to the Bradford law.There were 10 authors at least 2 published literatures, in who Bhasin DK in USA published 7 literatures, was the most active researcher in this field.The number of high frequency subject headings and subheadings was 33 and among which 5 research hotspots were clustered. ConclusionResearch hotspots about CT or MRI application in pancreatic pseudocyst mainly focuses on five aspects below:pathology, diagnosis, therapy, complications, and etiology.
Objective To learn the distribution pattern and worldwide research tendency of vitrectomy literatures. Methods Articles were searched from American Institute of Scientific Information (ISI) online database of web of science (WOS) database as a data source, to analyze the age distribution, national and regional, funding agency and citation of the vitrectomy literatures included during the year of 1971 -2011. The analysis software BibExcel and SPSS 19.0 were used to cluster highfrequency of them. Results Totally 8540 literatures were included, the numbers of them were gradually increased since 1971, significantly after 1991. The literatures were mainly in English, the literatures of our country capacity ranked 6th; funded institutions in all article, the National Natural Science Foundation of China ranked No. 5. Citation gradually increased since 1991, increased significantly after 2004. There were 50 highfrequency subjects, and hot topics were clustered into 6 categories which including vitrectomy for diseases of macula lutea, new techniques and complication of vitrectomy, medical treatment and surgical therapy of diabetic retinopathy, cataract, vitrectomy for endophthalmitis caused by intraocular injection and eye injury. Conclusions There is a growing trend on the research of vitrectomy. The hot topics include vitrectomy for diseases of macula lutea, new techniques and complication of vitrectomy. It may provide references for the scholars in scientific research and clinical studies.
Objective To investigate the hot topics of research on evidence-based medicine in 2002. Method To search MEDLINE for papers on evidence-based medicine published in 2002, identify high-frequency subject headings related to research on evidence-based medicine, and cluster the high-frequency subject headings according to rates they appeared in common papers. Results 545 papers, 1 554 subject headings, 30 high-frequency subject headings on evidence-based medicine, and nine clustering categories of high-frequency subject headings were identified through search. Conclusions Both high-frequency subject headings and their clustering categories suggested that “evidence-based practice guidelines and their innovation”, “evidence-based health research and health policy”, “methodology on systematic reviews and randomized clinical trials”, “method of evidence-based decision making and its application in various subjects”, were the hot topics of evidence-based medicine. They provided useful references for Chinese medical professionals to practice evidence-based medicine.
ObjectiveTo investigate the gene expression spectrum of retina and optic nerve after partial injury of optic nerve.MethodsSixty SD rats were randomly divided into 4 groups. The optic nerves of the right eyes were clipped for 6 seconds with a pair of crossaction forceps. The retinae and optic nerves in the operation eye and contralateral sham operation eye were removed 3, 7, 14, and 21 days after the injury to detect gene expression patterns with high-density DNA microarrays.ResultsChanges of a mass of gene expressions were found after the optic nerve injury, and the positive rate of gene expression was 2.35%, 6.48%, 3.82% and 4.09% after 3, 7, 14, 21 days, respectively, and the total positive rate was 11.77%. The functions of positive expression of the gene involved cell survival, cytoskeleton, extracellular matrix and cell adhesion, free radicals and oxidative damage, energy and metabolism, inflammation, neurotransmission and ion transport, signal transduction, structural protein, transcription and translation. Up-or down-regulation of repaired genes was the main part of the changes of gene expression, while the alteredexpression destroy genes was the minor part in the whole gene expression spectrum, in which the up- and down-regulation of expression of repaired genes accounted for 13.98% and 24.73% respectively 7 days after the injury, and the downregulation of expression of repaired genes accounted for 17.20% 14 days after the injury.ConclusionsA mass of gene expression changes occurs after the optic nerve injury, and the comprehensive view on the gene expression pattern following the optic nerve injury is crucial to discover the mechanism of post-injury reaction and regeneration.(Chin J Ocul Fundus Dis, 2005,21:163-166)
Objective To set up healthcare device-technology deployment assessment model and procedures through establishing the assessment parameter system between the functions of the clinical technical requirements and devices. Methods The bidirectional assessment parameter system developed by the literature review and Delphi, then combination weighting calculated by the combination weighting method, and the proposals for function deployment performed on the cluster analysis. Results The positive coefficients of twice Delphi were 75.56% and 87.50%, respectively. The effective recovery rates of the questionnaire were higher. The structure of the bidirectional assessment parameter system acquired according to the data mining and review, Delphi and integrated analysis. We calculated the weighting for the required functions and the deployed functions of the ventilator in the ICU, ER and RR. We listed the absolute importance and rank. The proposals for the function deployment of the ventilator which met different needs in fields of the critical care medicine were produced by the cluster analysis, ranking absolute importance and the calibration of weighting based on the investigation for actual function utilized rate. Conclusion It studies healthcare device-technology deployment assessment model by sequential integrated methods and sets up bidirectional assessment parameter system based on clinical technical function requirement, and the result is effective.
In the realm of data mining based on modern acupuncture clinical research, the impact of literature features such as literature quality, evidence level, sample size, and clinical efficacy on the quality of data mining outcomes remains uncertain. These issues are significant factors restricting the translational application of data mining research results. We suggest employing both entropy weight and linear weighting techniques to assess the specified indicators. This assessment results in a comprehensive weighted score for acupuncture prescriptions, serving as the foundation for our ensuing data mining endeavors. In this study, migraine research serves as an example to contrast the efficacy of weighted algorithms against that of classical algorithms. The findings demonstrate that the algorithm introduced in this research significantly contributes to studies focusing on the dispersed selection of acupuncture points. Its superiority lies in cluster analysis, where it adeptly discerns potential patterns in the amalgamation of acupoints. This algorithm amalgamates evidence-based acupuncture with data mining processes, providing innovative perspectives that augment the caliber of research in acupuncture data mining. Nonetheless, additional research is essential to corroborate these results.
This article analyzes the supply and demand data of outpatient resources in a large comprehensive tertiary grade A hospital from 2021 to 2023. Cluster analysis is used to classify the offline outpatient volume of each department and identify five different department categories with different outpatient volume characteristics. Based on the differences in outpatient volume and resource utilization between different categories and departments, this paper explores the supply-demand matching relationship of outpatient resources under normal and emergency states from online and offline outpatient. Based on the dimensions of categories and departments, this paper proposes an outpatient resource planning strategy that takes into account both normal and emergency states, providing a basis for improving the quality and efficiency of outpatient services in large comprehensive tertiary grade A hospitals.