Graph-based Clustering Method Research Articles

A new graph-based clustering method with application to single-cell RNA-seq data from human pancreatic islets.

Traditional bulk RNA-sequencing of human pancreatic islets mainly reflects transcriptional response of major cell types. Single-cell RNA sequencing technology enables transcriptional characterization of individual cells, and thus makes it possible to detect cell types and subtypes. To tackle the heterogeneity of single-cell RNA-seq data, powerful and appropriate clustering is required to facilitate the discovery of cell types. In this paper, we propose a new clustering framework based on a graph-based model with various types of dissimilarity measures. We take the compositional nature of single-cell RNA-seq data into account and employ log-ratio transformations. The practical merit of the proposed method is demonstrated through the application to the centered log-ratio-transformed single-cell RNA-seq data for human pancreatic islets. The practical merit is also demonstrated through comparisons with existing single-cell clustering methods. The R-package for the proposed method can be found at https://github.com/Zhang-Data-Science-Research-Lab/LrSClust.

Application of Multi-Resolution Graph-based Clustering for electrofacies prediction: a case study from the Horn River Shale, British Columbia, Canada

The Horn River Basin is a major shale gas play in British Columbia, Canada. An important component for identifying productive zones is to determine lithofacies. Quantitative core-based lithofacies classifications distinguish reservoirs and non-reservoir quality zones. In this study, Elemental capture spectroscopy (ECS) log data, XRD, and TOC data were integrated with core-description data to define the shale gas development interval in the Horn River Shale. Six lithofacies were determined and used as a training data for the supervised electrofacies classification. The Multi-Resolution Graph-based Clustering method (MRGC) was useful for prediction of electrofacies. The MRGC facilitated estimation of the electrofacies from the well logs of the non-coring wells based on our current construction model. We constructed a supervised MRGC model from well A. The probability scores of the electrofacies were 83%. This method was applied to non-coring wells B and C. Lateral variations in the facies were inferred using an electrofacies correlation or seismic-scale spatial model. Stochastic methods were applied to build a 3D facies model due to the insufficient electrofacies data. To increase the accuracy and applicability we used seismically derived-trend data (density and P-wave sonic inversion, envelope attributes) for sequential indicator simulation (SIS) modeling. After constructing a facies model, the faintly laminated siliceous mudstone (FLSM) and homogeneous siliceous mudstone (HSM) facies were shown to be dominantly distributed throughout the Evie Member and Muskwa Formation. These facies contain high silica compared to clay and higher TOC contents which are generally the main targets of a shale gas play. Therefore, the FLSM and HSM facies of the Muskwa Formation and Evie Member are potentially productive facies for shale gas development in the Horn River Shale.

A novel graph-based clustering method using noise cutting

Recently, many methods have appeared in the field of cluster analysis. Most existing clustering algorithms have considerable limitations in dealing with local and nonlinear data patterns. Algorithms based on graphs provide good results for this problem. However, some widely used graph-based clustering methods, such as spectral clustering algorithms, are sensitive to noise and outliers. In this paper, a cut-point clustering algorithm (CutPC) based on a natural neighbor graph is proposed. The CutPC method performs noise cutting when a cut-point value is above the critical value. Normally, the method can automatically identify clusters with arbitrary shapes and detect outliers without any prior knowledge or preparatory parameter settings. The user can also adjust a coefficient to adapt clustering solutions for particular problems better. Experimental results on various synthetic and real-world datasets demonstrate the obvious superiority of CutPC compared with k-means, DBSCAN, DPC, SC, and DCore.

A graph-based clustering method with special focus on hyperspectral imaging

A common trait of the more established clustering algorithms such as K-Means and HCA is their tendency to focus mainly on the bulk features of the data which causes minor features to be attributed to larger clusters. For hyperspectral imaging this has the consequence that substances which are covered by only a few pixels tend to be overlooked and thus cannot be separated. If small lateral features such as particles are the research objective this might be the reason why cluster analysis fails. Therefore we propose a novel graph-based clustering algorithm dubbed GBCC which is sensitive to small variations in data density and scales its clusters according to the underlying structures. The analysis of the proposed method covers a comparison to K-Means, DBSCAN and KNSC using a 2D artificial dataset. Further the method is evaluated on a multisensor image of atmospheric particulate matter composed of Raman and EDX data as well as an FTIR image of microplastics.

Applications of graph theory in district heat network analysis at national scale

District heat networks can make a significant contribution to decarbonising domestic heat supply and have been successfully applied in a number of countries. Traditionally, district heat networks have been analysed at the local or city district scale. Performing this analysis at national scale presents new methodological challenges. Efficient techniques are required to work with large datasets and enable rapid iteration of results. As district heating fundamentally involves networks, graph theory is highly relevant for providing well established analysis methods that can scale to very large datasets. Two applications of graph theory are presented for this work. The first presents a method for estimating minimum pipe lengths needed to connect buildings, from which demand densities may be derived. The second presents a graph-based clustering method for connecting distributed supplies and demand regions. These methods have been applied as Swiss national scale.

Outer-Points shaver: Robust graph-based clustering via node cutting

Graph-based clustering is an efficient method for identifying clusters in local and nonlinear data patterns. Among the existing methods, spectral clustering is one of the most prominent algorithms. However, this method is vulnerable to noise and outliers. This study proposes a robust graph-based clustering method that removes the data nodes of relatively low density. The proposed method calculates the pseudo-density from a similarity matrix, and reconstructs it using a sparse regularization model. In this process, noise and the outer points are determined and removed. Unlike previous edge cutting-based methods, the proposed method is robust to noise while detecting clusters because it cuts out irrelevant nodes. We use a simulation and real-world data to demonstrate the usefulness of the proposed method by comparing it to existing methods in terms of clustering accuracy and robustness to noisy data. The comparison results confirm that the proposed method outperforms the alternatives.

Comparative Cyto-molecular Analysis of Repetitive DNA Provides Insights into the Differential Genome Structure and Evolution of Five Cucumis Species

The genus Cucumis, includes the cucumber (2n = 14), melon (2n = 24), and other wild species, which is a good model for studying genome organization and evolution due to their variation in genome size and basic chromosome number. In this study, five Cucumis species with different geographical origins and basic chromosome numbers (i.e., C. sativus, C. hystrix, C. melo, C. anguria, and C. metuliferus) were used to identify and characterize the repetitive DNA in detail using a phylogenetic method. Comparative cyto-molecular genetic analysis of repetitive DNA was carried out using a graph-based clustering method, construction of Neighbor-Joining tree and fluorescence in situ hybridization (FISH). The results revealed that the five Cucumis species had differences in the repeat content of their genome, as well as in the composition of repetitive DNA and their genomic proportions. Three species from subgenus Melo showed a decreased tendency in both repeat types and genomic proportions, while two species from subgenus Cucumis also showed a decreased tendency in repeat types, but an increased tendency in genomic proportions. Phylogenic analysis of Cucumis Ty1/Copia, Ty3/Gypsy, and 45S rDNA ITS regions revealed that C. sativus, C. hystrix, and C. melo were closely related species, which C. sativus and C. hystrix were closer, while C. anguria and C. metuliferus were closer to each other and further from the other three species. Differential accumulation and elimination of different repeat types divergently shaped the genomic architecture of these five Cucumis species, contributing to the genome's evolution and diversification. Overall, these results enhance our understanding of the genomes of these five Cucumis species, and contribute to a more holistic view of genome evolution and phylogenetics of this genus.

FeatureLego: Volume Exploration Using Exhaustive Clustering of Super-Voxels

We present a volume exploration framework, FeatureLego, that uses a novel voxel clustering approach for efficient selection of semantic features. We partition the input volume into a set of compact super-voxels that represent the finest selection granularity. We then perform an exhaustive clustering of these super-voxels using a graph-based clustering method. Unlike the prevalent brute-force parameter sampling approaches, we propose an efficient algorithm to perform this exhaustive clustering. By computing an exhaustive set of clusters, we aim to capture as many boundaries as possible and ensure that the user has sufficient options for efficiently selecting semantically relevant features. Furthermore, we merge all the computed clusters into a single tree of meta-clusters that can be used for hierarchical exploration. We implement an intuitive user-interface to interactively explore volumes using our clustering approach. Finally, we show the effectiveness of our framework on multiple real-world datasets of different modalities.

GLASSgo - Automated and Reliable Detection of sRNA Homologs From a Single Input Sequence.

Bacterial small RNAs (sRNAs) are important post-transcriptional regulators of gene expression. The functional and evolutionary characterization of sRNAs requires the identification of homologs, which is frequently challenging due to their heterogeneity, short length and partly, little sequence conservation. We developed the GLobal Automatic Small RNA Search go (GLASSgo) algorithm to identify sRNA homologs in complex genomic databases starting from a single sequence. GLASSgo combines an iterative BLAST strategy with pairwise identity filtering and a graph-based clustering method that utilizes RNA secondary structure information. We tested the specificity, sensitivity and runtime of GLASSgo, BLAST and the combination RNAlien/cmsearch in a typical use case scenario on 40 bacterial sRNA families. The sensitivity of the tested methods was similar, while the specificity of GLASSgo and RNAlien/cmsearch was significantly higher than that of BLAST. GLASSgo was on average ∼87 times faster than RNAlien/cmsearch, and only ∼7.5 times slower than BLAST, which shows that GLASSgo optimizes the trade-off between speed and accuracy in the task of finding sRNA homologs. GLASSgo is fully automated, whereas BLAST often recovers only parts of homologs and RNAlien/cmsearch requires extensive additional bioinformatic work to get a comprehensive set of homologs. GLASSgo is available as an easy-to-use web server to find homologous sRNAs in large databases.

Diversity and evolution of the repetitive genomic content in Cannabis sativa

BackgroundThe repetitive content of the genome, once considered to be “junk DNA”, is in fact an essential component of genomic architecture and evolution. In this study, we used the genomes of three varieties of Cannabis sativa, three varieties of Humulus lupulus and one genotype of Morus notabilis to explore their repetitive content using a graph-based clustering method, designed to explore and compare repeat content in genomes that have not been fully assembled.ResultsThe repetitive content in the C. sativa genome is mainly composed of the retrotransposons LTR/Copia and LTR/Gypsy (14% and 14.8%, respectively), ribosomal DNA (2%), and low-complexity sequences (29%). We observed a recent copy number expansion in some transposable element families. Simple repeats and low complexity regions of the genome show higher intra and inter species variation.ConclusionsAs with other sequenced genomes, the repetitive content of C. sativa’s genome exhibits a wide range of evolutionary patterns. Some repeat types have patterns of diversity consistent with expansions followed by losses in copy number, while others may have expanded more slowly and reached a steady state. Still, other repetitive sequences, particularly ribosomal DNA (rDNA), show signs of concerted evolution playing a major role in homogenizing sequence variation.

DBH: A de Bruijn graph-based heuristic method for clustering large-scale 16S rRNA sequences into OTUs

Recent sequencing revolution driven by high-throughput technologies has led to rapid accumulation of 16S rRNA sequences for microbial communities. Clustering short sequences into operational taxonomic units (OTUs) is an initial crucial process in analyzing metagenomic data. Although many heuristic methods have been proposed for OTU inferences with low computational complexity, they just select one sequence as the seed for each cluster and the results are sensitive to the selected sequences that represent the clusters. To address this issue, we present a de Bruijn graph-based heuristic clustering method (DBH) for clustering massive 16S rRNA sequences into OTUs by introducing a novel seed selection strategy and greedy clustering approach. Compared with existing widely used methods on several simulated and real-life metagenomic datasets, the results show that DBH has higher clustering performance and low memory usage, facilitating the overestimation of OTUs number. DBH is more effective to handle large-scale metagenomic datasets. The DBH software can be freely downloaded from https://github.com/nwpu134/DBH.git for academic users.

Building an Urban Spatial Structure from Urban Land Use Data: An Example Using Automated Recognition of the City Centre

It has been suggested that the method of constructing an urban spatial structure typically follows a forward process from planning and design up to expression, as reflected in both graphic and text descriptions of urban planning. Although unorthodox, the original status structures can be extracted and constructed from an existing urban land-use map. This approach not only provides the methodological foundation for urban spatial structure evolution and allows for a comparative and quantitative analysis between the existing and planned conditions, but also lays a theoretical basis for failure in scientific decision making during the planning phase. This study attempts to achieve this by identifying the city centre (a typical element of the urban spatial structure) from urban land use data. The city centre is a special region consisting of several units with particular spatial information, including geometric attributes, topological attributes, and thematic attributes. In this paper, we develop a methodology to support the delineation of the city centre, considering these factors. First, using commercial land data, we characterise the city centre as units based on a series of indicators, including geometric and thematic attributes, and integrate them into a composite index of “urban centrality”; Second, a graph-based spatial clustering method that considers both topological proximity and attribute similarity is designed and used to identify the city centre. The precise boundary of the city centre is subsequently delimited using a shape reconstruction method based on the cluster results. Finally, we present a case study to demonstrate the effectiveness and practicability of the methodology.

Multi-resolution graph-based clustering analysis for lithofacies identification from well log data: Case study of intraplatform bank gas fields, Amu Darya Basin

In this study, we used the multi-resolution graph-based clustering (MRGC) method for determining the electrofacies (EF) and lithofacies (LF) from well log data obtained from the intraplatform bank gas fields located in the Amu Darya Basin. The MRGC could automatically determine the optimal number of clusters without prior knowledge about the structure or cluster numbers of the analyzed data set and allowed the users to control the level of detail actually needed to define the EF. Based on the LF identification and successful EF calibration using core data, an MRGC EF partition model including five clusters and a quantitative LF interpretation chart were constructed. The EF clusters 1 to 5 were interpreted as lagoon, anhydrite flat, interbank, low-energy bank, and high-energy bank, and the coincidence rate in the cored interval could reach 85%. We concluded that the MRGC could be accurately applied to predict the LF in non-cored but logged wells. Therefore, continuous EF clusters were partitioned and corresponding LF were interpreted, and the distribution and petrophysical characteristics of different LF were analyzed in the framework of sequence stratigraphy.

A comparative study of 3D FZI and electrofacies modeling using seismic attribute analysis and neural network technique: A case study of Cheshmeh-Khosh Oil field in Iran

Abstract Electrofacies are used to determine reservoir rock properties, especially permeability, to simulate fluid flow in porous media. These are determined based on classification of similar logs among different groups of logging data. Data classification is accomplished by different statistical analysis such as principal component analysis, cluster analysis and differential analysis. The aim of this study is to predict 3D FZI (flow zone index) and Electrofacies (EFACT) volumes from a large volume of 3D seismic data. This study is divided into two parts. In the first part of the study, in order to make the EFACT model, nuclear magnetic resonance (NMR) log parameters were employed for developing an Electrofacies diagram based on pore size distribution and porosity variations. Then, a graph-based clustering method, known as multi resolution graph-based clustering (MRGC), was employed to classify and obtain the optimum number of Electrofacies. Seismic attribute analysis was then applied to model each relaxation group in order to build the initial 3D model which was used to reach the final model by applying Probabilistic Neural Network (PNN). In the second part of the study, the FZI 3D model was created by multi attributes technique. Then, this model was improved by three different artificial intelligence systems including PNN, multilayer feed-forward network (MLFN) and radial basis function network (RBFN). Finally, models of FZI and EFACT were compared. Results obtained from this study revealed that the two models are in good agreement and PNN method is successful in modeling FZI and EFACT from 3D seismic data for which no Stoneley data or NMR log data are available. Moreover, they may be used to detect hydrocarbon-bearing zones and locate the exact place for producing wells for the future development plans. In addition, the result provides a geologically realistic spatial FZI and reservoir facies distribution which helps to understand the subsurface reservoirs heterogeneities in the study area.

The Constrained Laplacian Rank Algorithm for Graph-Based Clustering

Graph-based clustering methods perform clustering on a fixed input data graph. If this initial construction is of low quality then the resulting clustering may also be of low quality. Moreover, existing graph-based clustering methods require post-processing on the data graph to extract the clustering indicators. We address both of these drawbacks by allowing the data graph itself to be adjusted as part of the clustering procedure. In particular, our Constrained Laplacian Rank (CLR) method learns a graph with exactly k connected components (where k is the number of clusters). We develop two versions of this method, based upon the L1-norm and the L2-norm, which yield two new graph-based clustering objectives. We derive optimization algorithms to solve these objectives. Experimental results on synthetic datasets and real-world benchmark datasets exhibit the effectiveness of this new graph-based clustering method.

Reduction Techniques for Graph-Based Convex Clustering

The Graph-based Convex Clustering (GCC) method has gained increasing attention recently. The GCC method adopts a fused regularizer to learn the cluster centers and obtains a geometric clusterpath by varying the regularization parameter. One major limitation is that solving the GCC model is computationally expensive. In this paper, we develop efficient graph reduction techniques for the GCC model to eliminate edges, each of which corresponds to two data points from the same cluster, without solving the optimization problem in the GCC method, leading to improved computational efficiency. Specifically, two reduction techniques are proposed according to tree-based and cyclic-graph-based convex clustering methods separately. The proposed reduction techniques are appealing since they only need to scan the data once with negligibly additional cost and they are independent of solvers for the GCC method, making them capable of improving the efficiency of any existing solver. Experiments on both synthetic and real-world datasets show that our methods can largely improve the efficiency of the GCC model.

Summarizing video sequence using a graph-based hierarchical approach

Video summarization is a simplification of video content for compacting the video information. The video summarization problem can be transformed into a clustering problem, in which some frames are selected to saliently represent the video content. In this work, we use a graph-based hierarchical clustering method for computing a video summary. In fact, the proposed approach, called HSUMM, adopts a hierarchical clustering method to generate a weight map from the frame similarity graph in which the clusters (or connected components of the graph) can easily be inferred. Moreover, the use of this strategy allows the application of a similarity measure between clusters during graph partition, instead of considering only the similarity between isolated frames. We also provide a unified framework for video summarization based on minimum spanning tree and weight maps in which HSUMM could be seen as an instance that uses a minimum spanning tree of frames and a weight map based on hierarchical observation scales computed over that tree. Furthermore, a new evaluation measure that assesses the diversity of opinions among users when they produce a summary for the same video, called Covering, is also proposed. During tests, different strategies for the identification of summary size and for the selection of keyframes were analyzed. Experimental results provide quantitative and qualitative comparison between the new approach and other popular algorithms from the literature, showing that the new algorithm is robust. Concerning quality measures, HSUMM outperforms the compared methods regardless of the visual feature used in terms of F-measure.

Extraction of salient objects based on image clustering and saliency

Over the past decades, numerous methods have been proposed on salient object detection. However, most of these methods need users' interactions as a prerequisite to control their progress. In this paper, we propose a novel method for extraction of salient objects based on image clustering and saliency map from natural scene images. This method is a combination of image clustering, saliency map generation and automatic initialization. First, a graph based clustering method is applied to split the input image into regions. Second, a saliency map of the input image is generated using the contrast among split regions. From the split regions and generated saliency map, an adaptive threshold is defined, which classify the split regions into foreground and background. After that, the initial mask for object detection is determined using the classified foreground and background clusters and saliency values. A grab-cut with our initial mask is applied to extract the objects of interest, and the experimental results have shown that our proposed method is able to replace manual labeling of initialization in object detection.

Pore facies analysis: incorporation of rock properties into pore geometry based classes in a Permo-Triassic carbonate reservoir in the Persian Gulf

Pore facies analysis is a useful method for the classification of reservoir rocks according to pore geometry characteristics. The importance of this method is related to the dependence of the dynamic behaviour of the reservoir rock on the pore geometry. In this study, pore facies analysis was performed by the quantification and classification of the mercury injection capillary pressure (MICP) curves applying the multi-resolution graph-based clustering (MRGC) method. Each pore facies includes a limited variety of rock samples with different depositional fabrics and diagenetic histories, which are representative of one type of pore geometry. The present pore geometry is the result of the interaction between the primary rock fabric and its diagenetic overprint. Thus the variations in petrographic properties can be correlated with the pore geometry characteristics. Accordingly, the controlling parameters in the pore geometry characteristics were revealed by detailed petrographic analysis in each pore facies. The reservoir rock samples were then classified using the determined petrographic properties which control the pore system quality. This method is proposed for the classification of reservoir rocks in complicated carbonate reservoirs, in order to reduce the incompatibility of traditional facies analysis with pore system characteristics. The method is applicable where enough capillary pressure data is not available.

A Multidimensional and Multimembership Clustering Method for Social Networks and Its Application in Customer Relationship Management

Community detection in social networks plays an important role in cluster analysis. Many traditional techniques for one-dimensional problems have been proven inadequate for high-dimensional or mixed type datasets due to the data sparseness and attribute redundancy. In this paper we propose a graph-based clustering method for multidimensional datasets. This novel method has two distinguished features: nonbinary hierarchical tree and the multi-membership clusters. The nonbinary hierarchical tree clearly highlights meaningful clusters, while the multimembership feature may provide more useful service strategies. Experimental results on the customer relationship management confirm the effectiveness of the new method.