Distance-based Clustering Research Articles

Distance-based Clustering of Functional Data with Derivative Principal Component Analysis

Functional data analysis (FDA) is an important modern paradigm for handling infinite-dimensional data. An important task in FDA is clustering, which identifies subgroups based on the shapes of measured curves. Considering that derivatives can provide additional useful information about the shapes of functionals, we propose a novel L 2 distance between two random functions by incorporating the functions and their derivative information to determine the dissimilarity of curves under a unified scheme for dense observations. The Karhunen–Loève expansion is used to approximate the curves and their derivatives. Cluster membership prediction for each curve intends to minimize the new distances between the observed and predicted curves through subspace projection among all possible clusters. We provide consistent estimators for the curves, curve derivatives, and the proposed distance. Identifiability issues of the clustering procedure are also discussed. The utility of the proposed method is illustrated via simulation studies and applications to two real datasets. The proposed method can considerably improve cluster performance compared with existing functional clustering methods. Supplementary materials for the article are available online.

A strategy to detect metabolic changes induced by exposure to chemicals from large sets of condition-specific metabolic models computed with enumeration techniques

BackgroundThe growing abundance of in vitro omics data, coupled with the necessity to reduce animal testing in the safety assessment of chemical compounds and even eliminate it in the evaluation of cosmetics, highlights the need for adequate computational methodologies. Data from omics technologies allow the exploration of a wide range of biological processes, therefore providing a better understanding of mechanisms of action (MoA) related to chemical exposure in biological systems. However, the analysis of these large datasets remains difficult due to the complexity of modulations spanning multiple biological processes.ResultsTo address this, we propose a strategy to reduce information overload by computing, based on transcriptomics data, a comprehensive metabolic sub-network reflecting the metabolic impact of a chemical. The proposed strategy integrates transcriptomic data to a genome scale metabolic network through enumeration of condition-specific metabolic models hence translating transcriptomics data into reaction activity probabilities. Based on these results, a graph algorithm is applied to retrieve user readable sub-networks reflecting the possible metabolic MoA (mMoA) of chemicals. This strategy has been implemented as a three-step workflow. The first step consists in building cell condition-specific models reflecting the metabolic impact of each exposure condition while taking into account the diversity of possible optimal solutions with a partial enumeration algorithm. In a second step, we address the challenge of analyzing thousands of enumerated condition-specific networks by computing differentially activated reactions (DARs) between the two sets of enumerated possible condition-specific models. Finally, in the third step, DARs are grouped into clusters of functionally interconnected metabolic reactions, representing possible mMoA, using the distance-based clustering and subnetwork extraction method. The first part of the workflow was exemplified on eight molecules selected for their known human hepatotoxic outcomes associated with specific MoAs well described in the literature and for which we retrieved primary human hepatocytes transcriptomic data in Open TG-GATEs. Then, we further applied this strategy to more precisely model and visualize associated mMoA for two of these eight molecules (amiodarone and valproic acid). The approach proved to go beyond gene-based analysis by identifying mMoA when few genes are significantly differentially expressed (2 differentially expressed genes (DEGs) for amiodarone), bringing additional information from the network topology, or when very large number of genes were differentially expressed (5709 DEGs for valproic acid). In both cases, the results of our strategy well fitted evidence from the literature regarding known MoA. Beyond these confirmations, the workflow highlighted potential other unexplored mMoA.ConclusionThe proposed strategy allows toxicology experts to decipher which part of cellular metabolism is expected to be affected by the exposition to a given chemical. The approach originality resides in the combination of different metabolic modelling approaches (constraint based and graph modelling). The application to two model molecules shows the strong potential of the approach for interpretation and visual mining of complex omics in vitro data. The presented strategy is freely available as a python module (https://pypi.org/project/manamodeller/) and jupyter notebooks (https://github.com/LouisonF/MANA).

Clustering Uniformity Methods for Energy Efficiency in Wireless Sensor Networks

The wireless sensors that make up a wireless sensor network (WSN) are randomly deployed in nature and cannot be artificially replaced when their batteries are depleted. Failure of communication connection between wireless sensors causes continuous connection attempts, which results in a lot of power dissipation and shortens the lifetime of the WSN. In this paper, we propose to extend the lifetime of WSNs by limiting the appropriate distance between the cluster head (CH) node and the communicating sensor nodes (SNs) so that a group of clusters of appropriate size can be formed on a two-dimensional plane. To equalize cluster size, sensor nodes with the shortest distance communicate with each other to form member nodes, and nodes with closer distances are bring together to form clusters. The simulation results show the improvement rate of cluster uniformity over the shortest distance-based clustering method for clustering based on the proposed cluster uniformity algorithm. The proposed method can improve the cluster uniformity of the network by about 20%. In addition, the power consumption of the proposed method is analyzed according to the difference in the density of sensor nodes in the cluster groups to examine the improvement in power consumption.

Application of a hybrid-driven framework based on sensor optimization placement for the thermal error prediction of the spindle-bearing system

The precise thermal error prediction of spindle-bearing systems (SBSs) necessitates a comprehensive analysis of information gathered from multi-source sensors. However, limited data availability due to structural constraints poses challenges to fully characterize the system state. In this study, we introduce a data-model hybrid-driven framework based on sensor optimization placement for accurate thermal error prediction of SBSs. Firstly, a thermal hypernetwork method is developed to consider uneven temperature distribution and establish a unified information fusion model for state estimation. Secondly, based on an analysis of the rapidity and robustness, robust geodesic distance-based fuzzy c-medoid clustering with a simulated annealing algorithm (RGDFCMSA) is proposed to optimize sensor placement by minimizing the information entropy of the system. Next, uncertain parameters with estimability are selected based on SIAN and Sobol’s sensitivity indicator under optimal sensor placement. Furthermore, a multilayer particle filter (MLPF) is proposed to estimate temperature fields and predict the thermal error of SBSs by fusing information from multiple sources with different fidelity. Finally, experiments under different working conditions are conducted to validate the effectiveness and accuracy of the proposed method. The result indicates that the proposed framework is capable of an accurate estimation of the global temperature field, uncertain thermal parameters and thermal errors.

Six in one: cryptic species and a new host record for Olixon Cameron (Rhopalosomatidae, Hymenoptera) revealed by DNA barcoding

Olixon testaceum is a widely distributed species of brachypterous parasitoid wasp (Vespoidea: Rhopalosomatidae) occurring in Meso- and South America, but little is known of its biology. Here, the first known host of O. ? testaceum is identified as the cricket Anaxipha sp. (Grylloidea: Trigonidiidae) through DNA barcoding of six Olixon larvae and their hosts. Barcoding results also indicated substantial genetic diversity within nominal O. testaceum specimens. The number of species and statistical significance of these groups were tested using Maximum Likelihood phylogenies, distance-based cluster analyses, and coalescence models. All analyses revealed at least six distinct lineages, which suggests six or more cryptic species within O. ? testaceum. Combined with what is currently known about Rhopalosoma host use, these results indicate that rhopalosomatids may be generalist rather than specialist parasitoids, and further confirm the benefits of open global collaboration and DNA barcoding in advancing taxonomic knowledge.

Does spatial context of Third Place matter? Revisiting the Third Place concept based on user preferences: case of Kolkata, India

ABSTRACT Third Places (TPs), defined by Oldenburg in suburban American context are social informal gathering spaces, except home and workplace. Major literature focused on how TPs create social capital and cohesion within communities, largely benefitting socially, psychologically and even economically. Over the years, the TP concept is contested to globally transform from its initial theme of communitarianism to present dominant forms of neo-liberal consumerism. This posed significant threat to the inherent idea of affordable sociability rooted in characteristic plainness and neutral ground of TP. Moreover, exploring the spatial patterns and the influence of geographical variations on the nature of TP across urban landscapes is largely understudied. This study thus aimed to develop a consolidated understanding of the spatial dynamics of TPs in Indian context. The paper presented a city level, empirical based quantitative study based on online questionnaire survey. Commercial TPs are found to be most preferred. Research findings showed that the types of TP significantly affect the individual’s preferred travel distance. Also, one’s behavioural attributes as a whole significantly affect the catchment of the TPs, influenced by spatial variations. Hotspot analysis demonstrated three significant spatial concentrations across the city based on distance-based clustering. It showed heterogenous agglomerating tendencies of TP.

Genetic Polymorphisms and Forensic Parameters of Thirteen X-Chromosome Markers in the Iraqi Kurdish Population.

X-chromosome short tandem repeat (X-STR) tools are crucial in forensic genetics and human population fields. This study presents the development and validation of a multiplex STR system consisting of thirteen X-STR loci and amelogenin specific to the human X chromosome. The system was optimized and tested for species specificity, sensitivity, stability, and DNA mixture using 9947A female and 9948 male control genomic DNA. The amplified products of nine loci were sequenced to determine the correct amplicon length. Allele frequencies, forensic parameters, mean exclusion chance (MEC), linkage disequilibrium (LD), and allelic patterns were investigated using DNA samples from 225 (159 male, 66 female) unrelated Kurdish individuals who live in Sulaymaniyah province in the Kurdistan region of Iraq. The most informative locus in the Kurdish population was GATA172D05, while the least informative locus was DXS10164. The results demonstrated that the 13 X-STR system is highly polymorphic and sensitive for forensic DNA identification. Genetic distance-based clustering, metric multidimensional scaling (MDS), and correlation matrix were analyzed for 19 ethnic groups and populations. The phylogenetic tree showed that populations clustered according to their ethnogeographic relationships. The findings revealed genetic links between the Iraqi Kurds, Caucasians, Iraqi Arabs, United States (U.S.) ethnic groups, and Chinese populations.

Genetic variation and structure of endemic and endangered wild celery (Kelussia odoratissima Mozaff.) quantified using novel microsatellite markers developed by next-generation sequencing.

Kelussia odoratissima Mozaff. (Apiaceae) is a native plant that has been traditionally consumed in Iran's food and pharmaceutical industries. Overharvesting of the taxon, especially at the beginning of the growing season, due to its considerable medicinal and economic value, is believed to be the main reason for the extirpating of this plant. The consequences of the severe anthropogenic impacts on the genetic diversity of populations are poorly known. In order to investigate the level of genetic variation and patterns of the genetic structure of K. odoratissima, we developed novel microsatellite markers using the 454 Roche next-generation sequencing (NGS) platform for the first time. Out of 1,165 microsatellite markers bioinformatically confirmed, twenty-five were tested, of which 23 were used to screen genetic variation across 12 natural populations. Our results showed that the average number of alleles per locus and the polymorphic information content (PIC) were 10.87 (range 7 to 27), and 0.81 (range 0.67 to 0.94), respectively. The mean observed and expected heterozygosities (± SD) across all populations were 0.80 ± 0.31 and 0.72 ± 0.14, respectively. The average pairwise FST among the populations was 0.37 (range 0.04 to 0.81). Bayesian and distance-based clustering, and principal coordinate analyses revealed at least four major genetic clusters. Although high level of structure can be explained by landscape topography and geographic distance, presence of admixed populations can be associated to seed or pollen dispersal. Contrary to expectations, the high level of genetic variation and lack of inbreeding suggest that overexploitation has not yet significantly purged the allelic variability within the natural populations in protected areas.

Semi-supervised ensemble learning for human activity recognition in casas Kyoto dataset

—The automatic identification of human physical activities, commonly referred to as Human Activity Recognition (HAR), has garnered significant interest and application across various sectors, including entertainment, sports, and notably health. Within the realm of health, a myriad of applications exists, contingent upon the nature of experimentation, the activities under scrutiny, and the methodology employed for data and information acquisition. This diversity opens doors to multifaceted applications, including support for the well-being and safeguarding of elderly individuals afflicted with neurodegenerative diseases, especially in the context of smart homes. Within the existing literature, a multitude of datasets from both indoor and outdoor environments have surfaced, significantly contributing to the activity identification processes. One prominent dataset, the CASAS project developed by Washington State University (WSU) University, encompasses experiments conducted in indoor settings. This dataset facilitates the identification of a range of activities, such as cleaning, cooking, eating, washing hands, and even making phone calls. This article introduces a model founded on the principles of Semi-supervised Ensemble Learning, enabling the harnessing of the potential inherent in distance-based clustering analysis. This technique aids in the identification of distinct clusters, each encapsulating unique activity characteristics. These clusters serve as pivotal inputs for the subsequent classification process, which leverages supervised techniques. The outcomes of this approach exhibit great promise, as evidenced by the quality metrics' analysis, showcasing favorable results compared to the existing state-of-the-art methods. This integrated framework not only contributes to the field of HAR but also holds immense potential for enhancing the capabilities of smart homes and related applications.

Photovoltaic Solar Power Prediction Using iPSO-Based Data Clustering and AdaLSTM Network

Due to the increasing integration of photovoltaic (PV) solar power into power systems, the prediction of PV solar power output plays an important role in power system planning and management. This study combines an optimized data clustering method with a serially integrated AdaLSTM network to improve the accuracy and robustness of PV solar power prediction. During the data clustering process, the Euclidean distance-based clustering centroids are optimized by an improved particle swarm optimization (iPSO) algorithm. For each obtained data cluster, the AdaLSTM network is utilized for model training, in which multiple LSTMs are serially combined together through the AdaBoost algorithm. For PV power prediction tasks, the inputs of the testing set are classified into the nearest data cluster by the K-nearest neighbor (KNN) method, and then the corresponding AdaLSTM network of this cluster is used to perform the prediction. Case studies from two real PV stations are used for prediction performance evaluation. Results based on three prediction horizons (10, 30 and 60 min) demonstrate that the proposed model combining the optimized data clustering and AdaLSTM has higher prediction accuracy and robustness than other comparison models. The root mean square error (RMSE) of the proposed model is reduced, respectively, by 75.22%, 73.80%, 67.60%, 66.30%, and 64.85% compared with persistence, BPNN, CNN, LSTM, and AdaLSTM without clustering (Case A, 30 min prediction). Even compared with the model combining the K-means clustering and AdaLSTM, the RMSE can be reduced by 10.75%.

Identification of the mining accidents by a two-step clustering method for the mining-induced seismicity

Clustering methods aim to categorize data or samples into distinct groups based on their similarity. When applying clustering methods to earthquake events, it is crucial to establish a metric for quantifying the similarity between these events. Directly applying this clustering method to a catalog of mining-induced seismicity may lead to clustering earthquake events induced by different mining activities or accidents into the same group. To address this issue, a two-step clustering method has been proposed and applied for analyzing a catalog of mining-induced seismicity. The first step involves spatial distance-based clustering of seismic events, while the second step focuses on moment tensor analysis-based clustering of these events. The results obtained from the MT-based clustering method are visualized using Hudson Graphs, and box plots serve as an evaluation tool for assessing the quality of MT clustering. Most box plots demonstrate desirable quality in terms of MT cluster results, indicating successful outcomes. By the proposed two-step clustering method combined with actual mining activities, the potential accident locations and categories can be hypothesized while valuable recommendations provided for mining operations.

Label-aware distance mitigates temporal and spatial variability for clustering and visualization of single-cell gene expression data

Clustering and visualization are essential parts of single-cell gene expression data analysis. The Euclidean distance used in most distance-based methods is not optimal. The batch effect, i.e., the variability among samples gathered from different times, tissues, and patients, introduces large between-group distance and obscures the true identities of cells. To solve this problem, we introduce Label-Aware Distance (Lad), a metric using temporal/spatial locality of the batch effect to control for such factors. We validate Lad on simulated data as well as apply it to a mouse retina development dataset and a lung dataset. We also found the utility of our approach in understanding the progression of the Coronavirus Disease 2019 (COVID-19). Lad provides better cell embedding than state-of-the-art batch correction methods on longitudinal datasets. It can be used in distance-based clustering and visualization methods to combine the power of multiple samples to help make biological findings.

Impact of Dataset Scaling on Hierarchical Clustering: A Comparative Analysis of Distance-Based and Ratio-Based Methods

In this study, the distance-based agglomerative hierarchical clustering techniques were compared to a ratio-based approach. Two real datasets, which were also used in a prior study by Roux (2018), were considered. Firstly, it was observed that the type of scaling applied to the datasets was found to affect the results of hierarchical clustering. Thus, various scaling methods were employed prior to implementing hierarchical clustering. Furthermore, two rank-based goodness-of-fit measures were used to evaluate the hierarchical clustering methods. In contrast to Roux (2018) findings, it was observed that the distance-based methods, such as Median linkage, Average linkage, and centroid linkage, performed better than the ratio-based method. The ratio-based methods also showed issues with branch crossing in the hierarchical clustering dendrogram. Consequently, this study illustrates that, with appropriate dataset scaling, the distance-based methods outperform ratio-based methods in terms of goodness-of-fit measures.

The ensemble distance on model-based clustering for regions clustering based on rainfall: The case of rainfall in West Java Indonesia

Time series data clusters are being researched thoroughly. The distance metric drives the development of the clustering time series. The ARIMA model is one of the models that can be employed in model-based clustering, although differing model selection criteria can lead to uncertainty in the model. In this investigation, we created a technique for ensemble distance-based time series data clustering. To express the distance between two series, five distances based on the five model selection criteria are utilized. The average of the five distances reflects the distance of two time series data. According to the simulation results, the ensemble distance method could boost clustering accuracy by more than 11%. Based on the pattern of rainfall levels, we applied our methods to find clusters of locations in the Province of West Java (Indonesia). The findings indicate that the rainfall pattern in the same cluster is similar. The cluster model is effective and feasible for representing individual models in a cluster.

Adaptive cluster center initialization using density peaks for geodesic distance-based clustering

Adaptive cluster center initialization using density peaks for geodesic distance-based clustering

Genetic relatedness of rice landraces in North East India with wild relatives using chloroplast markers

AbstractThe evolutionary relationship between domesticated Oryza species and their wild relatives in North East India is not well understood. To improve the understanding of the evolutionary relationship, this study investigates the genetic diversity of 68 indigenous rice landraces from North East India, ten O. rufipogon genotypes, and nine O. nivara genotypes using chloroplast variable repeat markers which, due to non-recombination and uniparental inheritance, enable better phylogenetic inference. Reference genotypes IR64 (indica) and Nipponbare (japonica) were included to characterize various phylogenetic clusters. Using distance-based hierarchical clustering, model-based structuring and principal component analysis, selected landraces from the three North Eastern Indian states of Assam, Manipur and Arunachal Pradesh were grouped into two phylogenetically different clusters that represented the IR64 and Nipponbare groupings. Interestingly, despite the fact that a cluster analysis combining North East landraces and wild relatives likewise produced two separate clusters (cluster I: Nipponbare, cluster II: IR64), the majority of the wild relatives were only clustered in the IR64 cluster. This suggests that the two distinct evolutionary histories of the rice landraces in North East India are distinguished by their genetic affinity for wild relatives and their variation in the indica and japonica pools. These results highlight chloroplast divergence influencing the genetic diversity of North East landraces with wild relatives. Further, these findings will enable in-depth studies on the functional significance of chloroplast diversity on trait adaptation in rice landraces.

Parameter-free basis allocation for efficient multiple metric learning

Metric learning involves learning a metric function for distance measurement, which plays an important role in improving the performance of classification or similarity-based algorithms. Multiple metric learning is essential for efficiently reflecting the local properties between instances, as single metric learning has limitations in reflecting the nonlinear structure of complex datasets. Previous research has proposed a method for learning a smooth metric matrix function through data manifold to address the challenge of independently learning multiple metrics. However, this method uses the basic distance-based clustering algorithm to set the anchor points, which are the basis for local metric learning, and the number of basis metrics is dependent on the user. We propose a new method that can assign sophisticated anchor points by iteratively partitioning to identify mixed clusters of multi-class instances and cluster the most similar class instances together. In an experiment, we demonstrate the reliability of the automatically set parameter by comparison with the distribution of error rates according to the number of basis metrics of the existing algorithm. Furthermore, we show the superior performance of the proposed method over a fixed parameter setting of existing algorithms and confirm the relative classification accuracy superiority through performance comparison with baseline algorithms.

Functional Data Clustering Via Functional Mahalanobis Distance

As an exploratory data analysis method, functional data clustering aims to identify the underlying features of the observed data. In this context, this paper proposes a functional data clustering method based on functional Mahalanobis distance. As a distance-based non-parametric clustering model, the proposed method can effectively avoid the disadvantages of generative models and has excellent properties of decoupling and dimension standardization. Compared with other functional data clustering models, this method has lower computational complexity. In addition, the proposed method can be applied to any distance-based multivariate clustering method, thus generalizing it to the case of functional data. In practical data analysis, this paper compares the performance of this method with some other functional clustering methods, using k-means clustering as an example, and finds that it has better performance in terms of purity, adjusted Rand index, and computational speed. Finally, the idea of using Mahalanobis distance for functional data distance measurement can also be extended to construct kernel functions for measuring similarity between functional data samples, thus developing non-linear functional data analysis methods based on reproducing kernel theory.

Systematic analysis of plasmids of the Serratia marcescens complex using 142 closed genomes.

Plasmids play important roles in bacterial genome diversification. In the Serratia marcescens complex (SMC), a notable contribution of plasmids to genome diversification was also suggested by our recent analysis of >600 draft genomes. As accurate analyses of plasmids in draft genomes are difficult, in this study we analysed 142 closed genomes covering the entire complex, 67 of which were obtained in this study, and identified 132 plasmids (1.9-244.4 kb in length) in 77 strains. While the average numbers of plasmids in clinical and non-clinical strains showed no significant difference, strains belonging to clade 2 (one of the two hospital-adapted lineages) contained more plasmids than the others. Pangenome analysis revealed that of the 28 954 genes identified, 12.8 % were plasmid-specific, and 1.4 % were present in plasmids or chromosomes depending on the strain. In the latter group, while transposon-related genes were most prevalent (31.4 % of the function-predicted genes), genes related to antimicrobial resistance and heavy metal resistance accounted for a notable proportion (22.7 %). Mash distance-based clustering separated the 132 plasmids into 23 clusters and 50 singletons. Most clusters/singletons showed notably different GC contents compared to those of host chromosomes, suggesting their recent or relatively recent appearance in the SMC. Among the 23 clusters, 17 were found in only clinical or only non-clinical strains, suggesting the possible preference of their distribution on the environmental niches of host strains. Regarding the host strain phylogeny, 16 clusters were distributed in two or more clades, suggesting their interclade transmission. Moreover, for many plasmids, highly homologous plasmids were found in other species, indicating the broadness of their potential host ranges, beyond the genus, family, order, class or even phylum level. Importantly, highly homologous plasmids were most frequently found in Klebsiella pneumoniae and other species in the family Enterobacteriaceae, suggesting that this family, particularly K. pneumoniae, is the main source for plasmid exchanges with the SMC. These results highlight the power of closed genome-based analysis in the investigation of plasmids and provide important insights into the nature of plasmids distributed in the SMC.

The groundnut improvement network for Africa (GINA) germplasm collection: a unique genetic resource for breeding and gene discovery.

Cultivated peanut or groundnut (Arachis hypogaea L.) is a grain legume grown in many developing countries by smallholder farmers for food, feed, and/or income. The speciation of the cultivated species, that involved polyploidization followed by domestication, greatly reduced its variability at the DNA level. Mobilizing peanut diversity is a prerequisite for any breeding program for overcoming the main constraints that plague production and for increasing yield in farmer fields. In this study, the Groundnut Improvement Network for Africa assembled a collection of 1,049 peanut breeding lines, varieties, and landraces from 9 countries in Africa. The collection was genotyped with the Axiom_Arachis2 48K SNP array and 8,229 polymorphic single nucleotide polymorphism (SNP) markers were used to analyze the genetic structure of this collection and quantify the level of genetic diversity in each breeding program. A supervised model was developed using dapc to unambiguously assign 542, 35, and 172 genotypes to the Spanish, Valencia, and Virginia market types, respectively. Distance-based clustering of the collection showed a clear grouping structure according to subspecies and market types, with 73% of the genotypes classified as fastigiata and 27% as hypogaea subspecies. Using STRUCTURE, the global structuration was confirmed and showed that, at a minimum membership of 0.8, 76% of the varieties that were not assigned by dapc were actually admixed. This was particularly the case of most of the genotype of the Valencia subgroup that exhibited admixed genetic heritage. The results also showed that the geographic origin (i.e. East, Southern, and West Africa) did not strongly explain the genetic structure. The gene diversity managed by each breeding program, measured by the expected heterozygosity, ranged from 0.25 to 0.39, with the Niger breeding program having the lowest diversity mainly because only lines that belong to the fastigiata subspecies are used in this program. Finally, we developed a core collection composed of 300 accessions based on breeding traits and genetic diversity. This collection, which is composed of 205 genotypes of fastigiata subspecies (158 Spanish and 47 Valencia) and 95 genotypes of hypogaea subspecies (all Virginia), improves the genetic diversity of each individual breeding program and is, therefore, a unique resource for allele mining and breeding.