Results Of Hierarchical Clustering Research Articles

A novel procedure on next generation sequencing data analysis using text mining algorithm

BackgroundNext-generation sequencing (NGS) technologies have provided researchers with vast possibilities in various biological and biomedical research areas. Efficient data mining strategies are in high demand for large scale comparative and evolutional studies to be performed on the large amounts of data derived from NGS projects. Topic modeling is an active research field in machine learning and has been mainly used as an analytical tool to structure large textual corpora for data mining.MethodsWe report a novel procedure to analyse NGS data using topic modeling. It consists of four major procedures: NGS data retrieval, preprocessing, topic modeling, and data mining using Latent Dirichlet Allocation (LDA) topic outputs. The NGS data set of the Salmonella enterica strains were used as a case study to show the workflow of this procedure. The perplexity measurement of the topic numbers and the convergence efficiencies of Gibbs sampling were calculated and discussed for achieving the best result from the proposed procedure.ResultsThe output topics by LDA algorithms could be treated as features of Salmonella strains to accurately describe the genetic diversity of fliC gene in various serotypes. The results of a two-way hierarchical clustering and data matrix analysis on LDA-derived matrices successfully classified Salmonella serotypes based on the NGS data. The implementation of topic modeling in NGS data analysis procedure provides a new way to elucidate genetic information from NGS data, and identify the gene-phenotype relationships and biomarkers, especially in the era of biological and medical big data.ConclusionThe implementation of topic modeling in NGS data analysis provides a new way to elucidate genetic information from NGS data, and identify the gene-phenotype relationships and biomarkers, especially in the era of biological and medical big data.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-1075-9) contains supplementary material, which is available to authorized users.

Characterization of the duck (Anas platyrhynchos) Rbm24 and Rbm38 genes and their expression profiles in myoblast and skeletal muscle tissues.

RNA-binding motif proteins 24 (Rbm24) and 38 (Rbm38) are known to regulate genes expression in a post-transcriptional way. However, it remains unclear about the similarities and differences between Rbm24 and Rbm38 in terms of their sequence characteristics and expression profiles during myoblast differentiation and skeletal muscle development. In this study, we found that the coding domain sequences (CDSs) of duck Rbm24 and Rbm38 consisted of 678 and 648 nucleotides, respectively. Both of them contain a conserved RNA-recognition motif (RRM). Phylogenetic analysis showed that duck Rbm24 and Rbm38 were clustered with other Aves, nevertheless, avian Rbm24 was clustered with mammalian and reptilian Rbm24; avian Rbm38 was clustered with amphibian and reptilian Rbm38. Real-time PCR results exhibited that during embryonic stage, Rbm24 and Rbm38 in leg and breast muscle increased to their peak (P<0.01) at same time. During postnatal stage, the peak of Rbm24 and Rbm38 in leg were found at W5, while the peak of them in breast was found at W6 (P<0.01). Moreover, a relative high value of Rbm24 and Rbm38 in leg was found at W1 and W3, respectively. Additionally, both Rbm24 and Rbm38 expressed at each stage of duck myoblast differentiation, however, Rbm24 shared similar expression profiles with MEF2C, and Rbm38 shared similar expression profiles with MyoG and MEF2A. Furthermore, hierarchical clustering results were consistent with the preceding findings. These results serve as a foundation for further investigations about similar and different effects of Rbm24 and Rbm38 on myoblast differentiation and skeletal muscle development.

The Detection of Extremely High and Low Expressed Genes by EGEF Algorithm in Invasive Breast Cancer

Invasive breast cancer is a heterogeneous disease. The analysis of one or a group of specific gene expression profiles may not be enough to understand molecular activities in cancer cells. Therefore, a method which gives us the opportunity to compare similar up and down regulated gene expression profiles, is needed. The main purpose of our work is to sort the extreme high and low expressed genes and extract, compare and cluster them. Expression profiles of 598 samples of invasive breast cancer and 48 samples of normal tissue have been analysed to create a new algorithm called Extreme Gene Expression Family (EGEF). The EGEF algorithm sorted, grouped and compared the highest and the lowest expressed genes (n = 100). According to the hierarchical clustering result, dense and light memberships of gene families are detected. The resulting analysis allows us to predict which genes would show similar expression signatures in invasive breast cancer and to us to recognize specific biological activities and processes. EGEF algorithm can be used to detect expression signatures in other cancers and biological processes.

Recurrent Genetic Abnormalities Are Reflected in Phenotype at Diagnosis in Pediatric Acute Myeloid Leukemia: A Report from the Children's Oncology Group Protocol AAML0531

Background: Acute myeloid leukemia (AML) is a heterogeneous disease with high genomic and phenotypic complexity. Although a relationship between phenotype and t(15;17) is recognized, the effect of other genotypic abnormalities on phenotype remains to be elucidated.Objective: To explore the relationships between genotypic and phenotypic abnormalities in diagnostic specimens in pediatric AML.Methods: Of 1022 newly diagnosed pediatric patients with de novo AML enrolled on protocol AAML0531, 769 satisfied three criteria for this study: (1) submission of a blood or bone marrow sample for multidimensional flow cytometry (MDF) at diagnosis, (2) consent to blood bank specimen testing and (3) leukemia comprising >10% of non-erythroid cells by MDF. The diagnostic AML tumor population was identified by gating on CD45 vs log-SSC. A fifteen-dimensional immunophenotypic expression profile (IEP) was defined by computing mean fluorescent intensities, antigen intensity coefficient of variations, and light scatter characteristics of each leukemia without analyst imposed cut-offs.Unsupervised hierarchical clustering was performed to mathematically cluster patients with similar IEPs. The underlying karyotypic and mutational profile of each patient was compared to the IEPs to identify clusters of patients with relationships between phenotype and identifiable genetic abnormalities. Supervised bagged tree-based models were used to quantify the importance of each cell-surface antigen in identifying these clusters. Levene's test was used to assess the homogeneity of mean cell-surface antigen expression for important markers.Results: Analysis of hierarchical clustering results revealed 7 immunophenotypic clusters (A-G) of patients with relationships between IEP and 4 common genetic abnormalities (Figure 1).Cluster A and Cluster B were both strongly associated with the t(8;21) karyotype (Figure 1). 54/62 patients in Cluster A (87%) and 24/26 patients in Cluster B (92%) were positive for t(8;21). The most important cell-surface antigens in identifying patients with the Cluster A associated phenotype were dim CD33, expression of CD56, expression of CD34 and expression of HLA-DR. The mean antigen expression of these markers was significantly more homogenous for patients in Cluster A compared to all other patients for CD33 (p = 0.02), CD34 (p < 0.01) and HLA-DR (p<0.01). Mean expression of CD56 was notably more heterogeneous (p = 0.40) but positive in 91% of patients in this cluster. The most important cell-surface antigens in identifying the Cluster B associated phenotype were dim CD13, expression of CD34, expression of HLA-DR and dim CD33. The mean expression of these markers was significantly more homogeneous for patients in Cluster B compared to all other patients (all p values <0.01). In contrast to Cluster A, CD56 MFIs were negative for patients in Cluster B. Collectively, 72% of all patients with t(8;21) segregated into Cluster A (50%) or Cluster B (22%).Cluster C was strongly associated with chromosome 16 abnormalities (Figure 1). Of patients in this cluster, 50/55 (91%) were positive for inv(16)/t(16;16). The mean high-expression of CD13 was the most important antigen in identifying patients with this phenotype, and patients in Cluster C had more consistent mean CD13 expression compared to all other patients (p<0.01). 53% of all patients with chromosome 16 abnormalities comprised this cluster.Cluster D was strongly associated with 11q23 abnormalities (Figure 1). Of patients in this cluster, 105/135 (78%) were positive for 11q23 translocations. The most discriminating markers in identifying these patients were dim-to-negative CD34 and dim-to-negative CD117. 69% of all patients with 11q23 abnormalities comprised this cluster.Clusters E, F and G were associated with FLT3 ITD mutations (Figure 1). Of patients in these clusters, 44%, 49% and 53% were positive for FLT3 ITD mutations. Cumulatively, 63% of all patients with FLT3 ITD mutations were grouped into one of these three clusters.Conclusion: Hierarchical clustering algorithms identify IEPs associated with common genetic abnormalities in pediatric AML, providing insight into the consistent dysregulation of cell-surface gene product expression in patients with common genetic abnormalities. Further analysis may reveal that IEPs can identify patients otherwise classified as good or standard risk that will have a poor outcome. [Display omitted] DisclosuresVoigt:Hematologics Inc.: Employment. Brodersen:Hematologics Inc.: Employment. Menssen:Hematologics Inc.: Employment. Loken:Hematologics Inc.: Equity Ownership.

Tracking crop varieties using genotyping-by-sequencing markers: a case study using cassava (Manihot esculenta Crantz).

BackgroundAccurate identification of crop cultivars is crucial in assessing the impact of crop improvement research outputs. Two commonly used identification approaches, elicitation of variety names from farmer interviews and morphological plant descriptors, have inherent uncertainty levels. Genotyping-by-sequencing (GBS) was used in a case study as an alternative method to track released varieties in farmers’ fields, using cassava, a clonally propagated root crop widely grown in the tropics, and often disseminated through extension services and informal seed systems. A total of 917 accessions collected from 495 farming households across Ghana were genotyped at 56,489 SNP loci along with a “reference library” of 64 accessions of released varieties and popular landraces.ResultsAccurate cultivar identification and ancestry estimation was accomplished through two complementary clustering methods: (i) distance-based hierarchical clustering; and (ii) model-based maximum likelihood admixture analysis. Subsequently, 30 % of the identified accessions from farmers’ fields were matched to specific released varieties represented in the reference library. ADMIXTURE analysis revealed that the optimum number of major varieties was 11 and matched the hierarchical clustering results. The majority of the accessions (69 %) belonged purely to one of the 11 groups, while the remaining accessions showed two or more ancestries. Further analysis using subsets of SNP markers reproduced results obtained from the full-set of markers, suggesting that GBS can be done at higher DNA multiplexing, thereby reducing the costs of variety fingerprinting. A large proportion of discrepancy between genetically unique cultivars as identified by markers and variety names as elicited from farmers were observed. Clustering results from ADMIXTURE analysis was validated using the assumption-free Discriminant Analysis of Principal Components (DAPC) method.ConclusionWe show that genome-wide SNP markers from increasingly affordable GBS methods coupled with complementary cluster analysis is a powerful tool for fine-scale population structure analysis and variety identification. Moreover, the ancestry estimation provides a framework for quantifying the contribution of exotic germplasm or older improved varieties to the genetic background of contemporary improved cultivars.Electronic supplementary materialThe online version of this article (doi:10.1186/s12863-015-0273-1) contains supplementary material, which is available to authorized users.

Abstract 1993: Global correlation analysis of HOXC6 expression in ovarian cancer

Abstract Introduction: HOXC6 is a member of the HOX gene family known to be dysregulated in many cancers. We previously reported gene and protein down-regulation of HOXC6 in ovarian cancer tissue and serum of high-grade serous carcinoma patients. This study evaluates the relationship of the HOXC6 gene expression with other genes from transcriptional microarray analysis of 92 ovarian tissues. Methods: Transcriptional microarray data from ovarian tissue was analyzed from 67 whole tissue samples and 25 micro-dissected samples. The 25 samples consisted of 9 malignant tumor tissues, 9 normal epithelia (NE) tissues and 7 Benign Tumor tissues. Genes with significant direct or reverse correlations to HOXC6 were initially identified from the 67-sample set (FDR &amp;lt; 0.05). Correlated genes identified in the 67-sample set were then reanalyzed to reconfirm their significance against the HOXC6 in 25-sample set (FDR &amp;lt; 0.05). To verify validity of correlations in cancer, expression values of genes in this second group were compared between NE and malignant samples using ANOVA (FDR &amp;lt; 0.05). This analysis pipeline identified 116 genes, which were characterized by STRING-DB 9.1 for protein-protein interactions and GO enrichment analysis. 65 of the 116 genes with a fold change (FC) magnitude &amp;gt; 2.0 were used in unsupervised hierarchical clustering against both NE and malignant samples. Results: The absolute values of correlation coefficients (|r|) in 116 significant genes ranged from 0.43-0.81; 33 had (|r|) &amp;gt; 0.70, and 65 had |FC| &amp;gt; 2.0. Of note because of their known association to ovarian cancer, this analysis revealed PBX3, MAL2, TPD52, ZEB2, PTTG1, KLK7, and KLK8. GO enrichment analysis identified developmental process (36 genes), cell differentiation (34), anatomical structure development (36), and regulation of signaling (27) consistent with HOXC6 known functions (FDR &amp;lt; 0.05). Protein-protein interaction networks revealed SMAD3 to be the central node with the largest cluster of connectivity linked to other HOXC6 correlated genes including PAX8, ZEB2 and UCHL5. Unsupervised hierarchical clustering analysis clearly distinguished between NE obtained from non-tumorous and NE obtained from tumorous tissues. Discussion: Overall this analysis demonstrates a strong correlation between HOXC6 expression in ovarian tissue and multiple developmental genes. The hierarchical clustering results may indicate that effects of the reported genes are not limited to the tumor tissue and can be traced back to normal epithelial cells. Therefore, future study of large sample sets of micro-dissected malignant and nearby normal epithelia – to identify underlying developmental pathways in ovarian cancer – will increase understanding of growth mechanisms of ovarian tumors and their interaction with normal cells in the microenvironment. Our previous work along with these new findings could potentially be used for future study of identifying ovarian cancer specific biomarkers. Citation Format: Pourya Naderi Yeganeh, Zahra Bahrani-Mostafavi, Christine Richardson, David L. Tait, M. Taghi Mostafavi. Global correlation analysis of HOXC6 expression in ovarian cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1993. doi:10.1158/1538-7445.AM2015-1993

Optimized leaf ordering with class labels for hierarchical clustering.

Hierarchical clustering is extensively used in the bioinformatics community to analyze biomedical data. These data are often tagged with class labels, as e.g. disease subtypes or gene ontology (GO) terms. Heatmaps in connection with dendrograms are the common standard to visualize results of hierarchical clustering. The heatmap can be enriched by an additional color bar at the side, indicating for each instance in the data set to which class it belongs. In the ideal case, when the clustering matches perfectly with the classes, one would expect that instances from the same class cluster together and the color bar consists of well-separated color blocks without frequent alteration of colors (classes). But even in the case when instances from the same class cluster perfectly together, the dendrogram might not reflect this important aspect due to the fact that its representation is not unique. In this paper, we propose a leaf ordering algorithm for the dendrogram that preserving the hierarchical clustering result tries to group instances from the same class together. It is based on the concept of dynamic programming which can efficiently compute the optimal or nearly optimal order, consistent with the structure of the tree.

Preliminary evaluation of five elephant grass cultivars harvested at different time for sugar production

Five elephant grass cultivars, Pennisetum purpureum. cv. Huanan (Huanan), P. purpureum. cv. N51 (N51), P. purpureum. cv. Sumu No. 2 (Sumu-2), (Pennisetum americanum×P. purpureum)×P. purpureum cv. Guimu No. 1 (Guimu-1) and P. americanum cv. Tift23A×P. purpureum cv. Tift N51 (Hybrid Pennisetum), at three harvest stages were studied. With dilute sulfuric acid pretreatment followed by enzymatic hydrolysis, it is found that cellulose conversion of the five elephant grass cultivars harvested in August and September is higher than that harvested in October. The cellulose conversion for elephant grass cultivars harvested in August and September follows an order of Hybrid Pennisetum>Sumu-2>Huanan>Guimu-1>N51. This may be explained by the fact that lignification is gradually strengthened with time, inhibiting degradation of cellulose and hemicellulose. Moreover, cellulose conversions of Hybrid Pennisetum, Sumu-2 and Huanan harvested in August and September are higher based on hierarchical clustering results.

Multivariate Analysis of Grain Yield and Its Attributing Traits in Different Maize Hybrids Grown under Heat and Drought Stress.

This study was carried out to evaluate F1 single cross-maize hybrids in four crop growing seasons (2010–2012). Morphological traits and physiological parameters of twelve maize hybrids were evaluated (i) to construct seed yield equation and (ii) to determine grain yield attributing traits of well-performing maize genotype using a previously unexplored method of two-way hierarchical clustering. In seed yield predicting equation photosynthetic rate contributed the highest variation (46%). Principal component analysis data showed that investigated traits contributed up to 90.55% variation in dependent structure. From factor analysis, we found that factor 1 contributed 49.6% variation (P < 0.05) with primary important traits (i.e., number of leaves per plant, plant height, stem diameter, fresh leaves weight, leaf area, stomata conductance, substomata CO2 absorption rate, and photosynthetic rate). The results of two-way hierarchical clustering demonstrated that Cluster III had outperforming genotype H12 (Sultan × Soneri) along with its most closely related traits (photosynthetic rate, stomata conductance, substomata CO2 absorption rate, chlorophyll contents, leaf area, and fresh stem weight). Our data shows that H12 (Sultan × Soneri) possessed the highest grain yield per plant under environmentally stress conditions, which are most likely to exist in arid and semiarid climatic conditions, such as in Pakistan.

A Framework for Hierarchical Ensemble Clustering

Ensemble clustering, as an important extension of the clustering problem, refers to the problem of combining different (input) clusterings of a given dataset to generate a final (consensus) clustering that is a better fit in some sense than existing clusterings. Over the past few years, many ensemble clustering approaches have been developed. However, most of them are designed for partitional clustering methods, and few research efforts have been reported for ensemble hierarchical clustering methods. In this article, a hierarchical ensemble clustering framework that can naturally combine both partitional clustering and hierarchical clustering results is proposed. In addition, a novel method for learning the ultra-metric distance from the aggregated distance matrices and generating final hierarchical clustering with enhanced cluster separation is developed based on the ultra-metric distance for hierarchical clustering. We study three important problems: dendrogram description, dendrogram combination, and dendrogram selection. We develop two approaches for dendrogram selection based on tree distances, and we investigate various dendrogram distances for representing dendrograms. We provide a systematic empirical study of the ensemble hierarchical clustering problem. Experimental results demonstrate the effectiveness of our proposed approaches.

An information theoretic approach to hierarchical clustering combination

In Hierarchical clustering, a set of patterns are partitioned into a sequence of groups represented as a dendrogram. The dendrogram is a tree representation where each node is associated with merging of two (or more) partitions and hence each partition is nested into the next partition. Hierarchical representation has properties that are useful for visualization and interpretation of clustering results. On one hand, different hierarchical clustering algorithms usually produce different dendrograms. On the other hand, clustering combination methods have received considerable interest in recent years and they yield superior results for clustering problems. This paper proposes a framework for combining various hierarchical clustering results which preserves the structural contents of input hierarchies. In this method, first a description matrix is created for each hierarchy, and then the description matrices of the input hierarchies are aggregated to form a consensus matrix from which the final hierarchy is derived. In this framework, we use two new measures for aggregating the description matrices, namely Rényi and Jensen–Shannon Divergences. The experimental and comparative analysis of our proposed framework shows the effectiveness of these two aggregators in hierarchical clustering combination.

Comparing weighted combination of hierarchical clustering based on Cophenetic measure

In Hierarchical Clustering (HC), a set of patterns is partitioned into a sequence of groups and the results are repre- sented by a dendrogram. The dendrogram is a tree representation where each node is associated with merging of two (or more) partitions and hence each partition is nested into the next partition. Hierarchical representation has valuable properties that are useful for visualization and interpretation of the clustering results. On one hand, it is well known that hierarchical clustering results are sensitive to the algorithms that are used, thus different dendrograms are created. On the other hand, clustering com- bination methods have demonstrated promising results. This motivated us to consider different combination method for a set of dendrograms. In this paper, we compare several weighted combination methods in which the hierarchical clustering results are used in order to derive a consensus hierarchical clustering. The Cophenetic measures are used for comparison. The results illustrate that the proposed weighing combination of hierarchical clusterings perform better than other combinations such as averaging.

Hierarchical hesitant fuzzy K-means clustering algorithm

Due to the limitation and hesitation in one’s knowledge, the membership degree of an element to a given set usually has a few different values, in which the conventional fuzzy sets are invalid. Hesitant fuzzy sets are a powerful tool to treat this case. The present paper focuses on investigating the clustering technique for hesitant fuzzy sets based on the K-means clustering algorithm which takes the results of hierarchical clustering as the initial clusters. Finally, two examples demonstrate the validity of our algorithm.

Hemodynamic Segmentation of Brain Perfusion Images with Delay and Dispersion Effects Using an Expectation-Maximization Algorithm

Automatic identification of various perfusion compartments from dynamic susceptibility contrast magnetic resonance brain images can assist in clinical diagnosis and treatment of cerebrovascular diseases. The principle of segmentation methods was based on the clustering of bolus transit-time profiles to discern areas of different tissues. However, the cerebrovascular diseases may result in a delayed and dispersed local perfusion and therefore alter the hemodynamic signal profiles. Assessing the accuracy of the segmentation technique under delayed/dispersed circumstance is critical to accurately evaluate the severity of the vascular disease. In this study, we improved the segmentation method of expectation-maximization algorithm by using the results of hierarchical clustering on whitened perfusion data as initial parameters for a mixture of multivariate Gaussians model. In addition, Monte Carlo simulations were conducted to evaluate the performance of proposed method under different levels of delay, dispersion, and noise of signal profiles in tissue segmentation. The proposed method was used to classify brain tissue types using perfusion data from five normal participants, a patient with unilateral stenosis of the internal carotid artery, and a patient with moyamoya disease. Our results showed that the normal, delayed or dispersed hemodynamics can be well differentiated for patients, and therefore the local arterial input function for impaired tissues can be recognized to minimize the error when estimating the cerebral blood flow. Furthermore, the tissue in the risk of infarct and the tissue with or without the complementary blood supply from the communicating arteries can be identified.

Comparison of compositional characteristics of amino acids between livestock wastewater and carcass leachate

This study was purposed to examine the use of amino acids as an indicator to determine whether groundwater around carcass burial sites is polluted by livestock wastewater (LW) or carcass leachate (CL). The analysis of samples showed that the average amino acid concentration of carcass leachate (531.897 mg/L; 4341.784 μmol/L) was about 300 times as high as that of livestock wastewater (1.755 mg/L; 16.283 μmol/L). To identify distinct characteristics between LW and CL, six amino acids were paired with one another to calculate their relative composition ratios, which were found to be Leu/Trp (CL 8.39∼98.6, LW 0.89∼4.77), Val/Trp (CL 11.95∼175.38, LW 0.73∼3.62), Lys/Leu (CL 0.01∼0.72, LW 0.96∼8.44), Lys/Ile (CL 0.02∼1.55, LW 1.64∼10.99), Met/Lys (CL 0.14∼0.45, LW 0.03∼0.14), and Ile/Val (CL 0.38∼0.73, LW 0.40∼0.97). The hierarchical clustering result showed that the similarity was 0.617 among the seven LW samples and 0.563 among the seven CL samples, while the similarity between LW and CL samples was 0.198, presenting that these two sources are distinct from each other. All these results indicate that amino acids can be used as a tracer to evaluate if the contamination source is livestock wastewater or carcass leachate. To apply amino acids to tracing pollutants more effectively, however, further studies are needed to understand whether the relative abundance ratios of amino acids are maintained as they are transporting through soils as a medium.

一种基于同步动力学模型的层次聚类方法

Clustering is an essential method for analyzing and mining the intrinsic group in data. This paper presents a novel synchronization-based hierarchical clustering method based on an extended Kuramoto dynamic synchronization model. Each data object is regarded as a phase oscillator and interacts dynamically with its neighboring objects. As time evolves, objects synchronize naturally. With regard to the local diameter of the neighborhood, the proposed method nds local synchronization-based natural clusters. Hierarchical clustering results are achieved by enlarging the local neighborhood distance of objects synchronizing continuously. Using a neighborhood closure, our method predicts clusters before the objects reach local synchronization, thereby signi cantly reducing the dynamic interaction time. To select the optimal clusters automatically, this hierarchical clustering method based on a dynamic synchronization model is combined with a clustering validation measure known as the silhouette width criterion. Combined with the silhouette width criterion, the proposed method is parameter-free. Moreover, the proposed method can detect clusters in data of arbitrary shapes, sizes and numbers without any data distribution assumptions. This synchronization-based clustering also allows natural outlier identi cation, since outliers do not synchronize with data objects in clusters. Extensive experiments on several synthetic and real-world data sets demonstrate that the proposed method achieves high clustering accuracy with lower execution time and fewer synchronization steps compared to the state-of-the-art method.

OP0076 Difference in the Expression of Long Intergenic Non-Coding RNA Between Polymyalgia Rheumatica and Rheumatoid Arthritis Might Be Important for Understanding and Discriminating these Diseases

BackgroundIt has been reported that the specificity for discriminating polymyalgia rheumatica (PMR) from rheumatoid arthritis (RA) was up to 70% with 2012 EULAR/ACR provisional classification criteria. There are many similarities...

Interactive visual summaries for detection and assessment of spatiotemporal patterns in geospatial time series

Numerous measurement devices and computer simulations produce geospatial time series that describe a wide variety of processes of System Earth. A major challenge in the analysis of such data is the complexity of the described processes, which requires a simultaneous assessment of the data’s spatial and temporal variability. To address this task, geoscientists often use automated analyses to compute a compact description of the data, ideally comprising characteristic spatial states of the process under study and their occurrence over time. The results of such automated methods depend on the parameterization, especially the number of extracted spatial states. A particular number of spatial states, however, may only reflect certain spatial or temporal aspects. We introduce a visual analytics approach that overcomes this limitation by allowing users to extract and explore various sets of spatial states to detect characteristic spatiotemporal patterns. To this end, we use the results of hierarchical clustering as a starting point. It groups all time steps of a geospatial time series into a hierarchy of clusters. Users can interactively explore this hierarchy to derive various sets of spatial states. To facilitate detailed inspection of these sets, we employ the concept of interactive visual summaries. A visual summary is the depiction of a set of spatial states and their associated time steps or intervals. It includes interactive means that allow users to assess how well the depicted patterns characterize the original data. Our visual interface comprises a system of visualization components to facilitate both the extraction of sets of spatial states from the hierarchical clustering output and their detailed inspection using interactive visual summaries. This study results from a close collaboration with geoscientists. In an exemplary analysis of observational ocean data, we show how our approach can help geoscientists gain a better understanding of geospatial time series.

Hierarchical clustering as a rapid tool for surveillance of emerging antibiotic-resistance phenotypes in Klebsiella pneumoniae strains

Antimicrobial resistance is on the rise, and its early detection and surveillance are critical to implement effective control measures. The aim of this study was to develop a rapid hierarchical clustering bioinformatic tool for application on antibiotic susceptibility testing (AST) results (resistant, intermediate, sensitive) of a series of Klebsiella pneumoniae clinical isolates from Algeria and from France for surveillance of antibiotic-resistance phenotypes. A total of 1011 K. pneumoniae strains were collected from August 2008 to December 2012: 221 clinical isolates from western Algeria and 790 clinical isolates from Marseille, France. AST against a panel of 16 antibiotics was done for all isolates. Results of AST were introduced into MultiExperiment Viewer (MeV) software to perform hierarchical clustering, with resistant, intermediate and sensitive being translated to 1, 0 and -1 values, respectively. Hierarchical clustering results were compared to standard resistance phenotypes to evaluate the accuracy of the method. Based on the AST results, the 221 K. pneumoniae strains from Algeria could be separated into six phenotype groups as regards their resistance to β-lactam compounds: extended spectrum β-lactamase (ESBL) (68.3 %), ESBL associated with cephalosporinase (13.1 %), cephalosporinase (0.9 %), penicillinase (3.6 %) and wild-type (14.0 %). Hierarchical clustering by the MeV software applied to the AST results for all 1011 isolates generated clusters that were significantly representative of phenotypic classification and geographical origin, in less than 1 min. Moreover, adding to the dataset the AST results of a K. pneumoniae NDM-1 positive strain, the only strain resistant to imipenem in the series, immediately generated a new branch in the dendrogram. We have developed a rapid and simple hierarchical clustering tool for application on AST results that was able to survey qualitatively and quantitatively the prevalence of known and unknown phenotypes. This tool could be easily implemented in routine clinical microbiology laboratories.

A heuristic hierarchical clustering based on multiple similarity measurements

Similarity is the core problem of clustering. Clustering algorithms that are based on a certain, fixed type of similarity are not sufficient to explore complicated structures. In this paper, a constructing method for multiple similarity is proposed to deal with complicated structures of data sets. Multiple similarity derives from the local modification of the initial similarity, based on the feedback information of elementary clusters. Combined with the proposed algorithm, the repeated modifications of local similarity measurement generate a hierarchical clustering result. Some synthetic and real data sets are employed to exhibit the superiority of the new clustering algorithm.