Hierarchical Clustering Model Research Articles

Molecular Characterization of Salmonella Typhimurium and E. coli O157:H7 Isolated from Ready-to-eat Chicken Meat

Background: Virulence genes are important in the pathogenesis of bacteria and in the mechanism of bacterial pathogenicity.
 Objective: The study aimed at molecular detection of virulence genes such as sdiA, fimH, InvA and fliC and determination of similarity and links among the isolates.
 Method: A total of 67 isolates including 4 controls were tested by Polymerase Chain Reaction (PCR) with 4 primer pairs including invA, fliC, sdiA and fimH to avoid bias. Full sequences of the 16S-Sequencing gene of both strains were carried out with 29 Salmonella and 30 E. coli isolates that were positive for amplification at 1500bp.ward hierarchical clustering model and agglomeration procedure was used. Clustered grouping and relational affinity test were conducted and depicted by the dendrogram. Molecular identification and interpretation were done using BLASTn Protocol.
 Result: None of the isolates were positive to the invA or fliC gene fragments. One isolate from each of Salmonella Typhimurium and E. coli O157:H7 was positive to sdiA and fimH respectively. Three E. coli isolates were positive with an amplification of 500bp which is specific for fimH genes. One of the isolates E459 showed amplification of fimH gene with multidrug resistance to 5 drugs namely Ceftazidime, Cefuroxime, Augmentin, Nitrofurantoin and Ampicillin. Four Salmonella isolates had an amplification of 274bp specific for sdiA gene. Two variants of E. coli O157:H7 (unit g1 and unit g2) were identified. A mutant strain Salmonella Typhimurium LTS (STMD1) causing human gastroenteritis was identified also, Salmonella Typhimurium DT104 isolated exhibited multiple resistant genes (ACSSuT, SGI1) against several antibiotics. These are of public health significance.
 Conclusion: This study has indicated presence of sdiA and fimH genes in Salmonella typhimurium and E. coli 0157:H7 respectively isolated from ready-to-eat chicken meats from public eateries. It also indicated association of sdiA and fimH genes with multi-drugs resistance.

Feature reduction for imbalanced data classification using similarity-based feature clustering with adaptive weighted K-nearest neighbors

Most existing imbalanced data classification models mainly focus on the classification performance of majority class samples, and many clustering algorithms need to manually specify the initial cluster centers and the number of clusters. To solve these drawbacks, this study presents a novel feature reduction method for imbalanced data classification using similarity-based feature clustering with adaptive weighted k-nearest neighbors (AWKNN). First, the similarity between samples is evaluated by the difference and smaller value between samples on each dimension, a similarity measure matrix is then developed to measure the similarity between clusters, after which a new hierarchical clustering model is constructed. By combining the cluster center of each sample cluster with its nearest neighbor, new samples are generated. Then, a hybrid sampling model based on similarity measure is presented by putting the generated samples into imbalanced data and removing samples from majority classes. Thus, a balanced decision system is constructed based on generated samples and minority class samples. Second, to address the issues that the traditional symmetric uncertainty only considers the correlation between features, and mutual information ignores the added information after classification, the normalized information gain is introduced to design new symmetric uncertainty between each feature and the other features; then, the ordered sequence and the average of the symmetric uncertainty difference of each feature are provided to adaptively select the k-nearest neighbors of features. Moreover, the weight of the k-th nearest neighbor of features is defined to present the AWKNN density of features and their ordered sequence for clustering features. Finally, by combining the weighted average redundancy with the symmetric uncertainty between features and decision classes, the maximum relevance between each feature and decision classes, and the minimum redundancy among features in the same cluster is presented to select the optimal feature subset from the feature clusters. Experiments applied to 29 imbalanced datasets show that the developed algorithm is effective and can select the optimal feature subset with high classification accuracy for imbalanced data.

Does clustering of DNA barcodes agree with botanical classification directly at high taxonomic levels? Trees in French Guiana as a case study.

Characterizing biodiversity is one of the main challenges for the coming decades. Most diversity has not been morphologically described, and barcoding is now complementing morphological-based taxonomy to further develop inventories. Both approaches have been cross-validated at the level of species and OTUs. However, many known species are not listed in reference databases. One path to speed up inventories using barcoding is to directly identify individuals at coarser taxonomic levels. We therefore studied in barcoding of plants whether morphological-based and molecular-based approaches are in agreement at genus, family and order levels. We used Agglomerative Hierarchical Clustering (with Ward, Complete and Single Linkage) and Stochastic Block Models (SBM), with two dissimilarity measures (Smith-Waterman scores, kmers). The agreement between morphological-based and molecular-based classifications ranges in most of the cases from good to very good at taxonomic levels above species, even though it decreases when taxonomic levels increase, or when using the tetramer-based distance. Agreement is correlated with the entropy of morphological-based classification and with the ratio of the mean within- and mean between-groups dissimilarities. The Ward method globally leads to the best agreement, whereas Single Linkage can show poor behaviours. SBM provides a useful tool to test whether or not the dissimilarities are structured by the botanical groups. These results suggest that automatic clustering and group identification at taxonomic levels above species are possible in barcoding.

Impacts of COVID-19 on Electric Vehicle Charging Behavior: Data Analytics, Visualization, and Clustering

COVID-19 pandemic has infected millions and led to a catastrophic loss of lives globally. It has also significantly disrupted the movement of people, businesses, and industries. Additionally, electric vehicle (EV) users have faced challenges in charging their vehicles in public charging locations where there is a risk of COVID-19 exposure. However, a case study of EV charging behavior and its impacts during the SARS-CoV-2 is not addressed in the existing literature. This paper investigates the impacts of COVID-19 on EV charging behavior by analyzing the charging activity during the pandemic using a dataset from a public charging facility in the USA. Data visualization of charging behavior alongside significant timelines of the pandemic was utilized for analysis. Moreover, a cluster analysis using k-means, hierarchical clustering, and Gaussian mixture models was performed to identify common groups of charging behavior based on the vehicle arrival and departure times. Although the number of vehicles using the charging station was reduced significantly due to lockdown restrictions, the charging activity started to pick up again since May 2021 due to an increase in vaccination and easing of public restrictions. However, the charging activity currently still remains around half of the activity pre-pandemic. A noticeable decline in charging session length and an increase in energy consumption can be observed as well. Clustering algorithms identified three groups of charging behavior during the pandemic and their analysis and performance comparison using internal validation measures were also presented.

Ontology-based mapping approach for automatic work packaging in modular construction

Many cross-knowledge domain tasks involving various professional backgrounds have been transferred from construction sites to factories in modular construction (MC). In MC, forming optimal work packages which can handle the complexity of product breakdown structures and dynamic project progress is critical for task planning and execution. However, forming MC work packages is time-consuming and ineffective because it is performed manually while not adequately considering domain knowledge. To address the problem, this study proposes a dynamic ontology-based mapping (DOM) approach to automatically generate semantic-enriched work packages. For this purpose, ontologies of MC products, topology, and tasks are established to incorporate domain knowledge. Then, a customized Latent Dirichlet Allocation (LDA) model for mapping products to tasks and a weighted hierarchical clustering model for grouping dynamic tasks into work packages are developed. The effectiveness of the DOM approach is tested in an MC case project and controlled experiments. The results demonstrate that the DOM approach can significantly increase the accuracy and efficiency of the dynamic work packaging process while reducing planning time compared to conventional methods, which thus improve the collaborative management and performance of MC projects.

Protein Expression Signatures Predict Response, Relapse, and Survival after Venetoclax Containing Therapy in Adults with Acute Myeloid Leukemia

IntroductionWhen added to cytarabine (Ara-c) or hypomethylating agents (HMA), the BCL2 inhibitor, venetoclax (VTX), has been reported to improve response and overall survival (OS) rates. However, resistance and relapse still occur in the majority, and, although alterations in MCL1 and BCLXL are noted at relapse, identification of prognostic features remain unknown, notably not correlating with expression of the BCL2 target. Identification of prognostic markers could guide VTX use in patients and/or post-remission therapy. We searched for protein expression targets individually and collectively to predict VTX response and relapse in AML.MethodsReverse Phase Protein Array (RPPA) was performed on diagnostic leukemia samples of 818 adults with AML, of which 143 received VTX including 33 in combination with high dose Ara-C, 5 with standard dose Ara-C, 50 with HMA, and 13 with HMA and targeted therapy. Protein expression levels were evaluated using 390 validated antibodies were analyzed in the context of clinical data compiled by retrospective chart review. Pearson correlation was used to identify significant protein-protein correlations. Survival curves were generated by the Kaplan-Meier method and survival data was analyzed by multivariate cox regression model. Protein expression signatures were identified by hierarchical clustering and predictive models of classifiers were determined by classification and regression trees (CART) analysis.ResultsWe queried the 390 proteins assayed in the 143 VTX treated patients to identify proteins individually prognostic (p<0.01) for OS (n=27) or remission duration (RD, n=44). Notably, neither MCL1, BCLXL nor BCL2 expression at diagnosis were prognostic of OS or RD. From these, unbiased hierarchical clustering revealed two cohorts (N=102 & 41 patients) for OS and RD. The clusters were similar for clinical features with no significant differences noted for, age, gender, performance status, cytogenetics, or the presence of molecular mutation markers FLT3.ITD, IDH1/2, NPM1 or TP53. The groups did not differ by therapy combination. Remission rates were insignificantly less in cluster 1 (61% vs 77%). Clear differences were observed for OS with estimated 3-yr overall survival 27% vs 66% (p=0.009, Figure 1) and relapse risk (RR) at 1-yr 45% vs 18% (p=0.001, Figure 2) in cluster 1 vs 2, respectively. In multivariate analysis, protein cluster membership was in independent prognostic factor for OS (along with TP53 and NPM1 mutations) but unfavorable cytogenetics was not. Prognostication did not vary based on cytogenetics or therapy received. For RD, protein cluster membership and unfavorable cytogenetics were the only independent predictors. Of the 44 proteins in the protein signature, CART modeling identified 3 - SPI1, NOTCH1.cle, and PTPN12 - that could predict clustering with a computed accuracy of 94.3%. Similarly, when these three proteins were used as training variables for random forest classification, the error rate was 3.7%. Several previously unrecognized potential therapeutic targets for preventing VTX resistance were also identified.DiscussionProtein expression patterns, individually and in combination, were very highly predictive of outcome to VTX containing combination chemotherapy. A group with lower response rates, higher relapse rates, shorter RD and inferior OS was defined. A kit to prospectively determine cluster membership is in development. If validated this could be used to triage high-risk patients to alternate therapies, such as transplant, in CR1. Many new targets for combination therapy to prevent VTX resistance were identified and need to be tested in the laboratory for clinical relevance. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

Site Fidelity and Residency of Tursiops truncatus off the Aragua Coast, Venezuela-First Records of Long Residency

This study represents the first comprehensive analysis of the residency patterns of a coastal population of bottlenose dolphin off the coast of Aragua, Venezuela, over a multi-year period. Using photo-identification, the most recent study (2019-2020) identified 56 individuals with the time between encounters from one to 344 days between the first and last sighting. Site Fidelity (SF) and Residence (RES) indices were calculated and Agglomerative Hierarchical Clustering (AHC) modeling was performed, with three patterns of residence obtained: resident (25%), semiresident (17.86%) and transient (57.14%). These results were contrasted with remodeled data from a previous study (2006-2007), showing similar patterns: resident (24.44%), semi-resident (28.89%) and transient (46.67%). Importantly, two individuals were found to have been resident over the extended period. A breeding female sighted for the first time in 2004 and again in 2020 (16 years) and the other from 2005 to 2020 (15 years). This region is an important area for marine mammals, known to support a resident reproductive population over many years, as well seabirds, sea turtles, whale sharks and fishermen. We recommend that consideration be given to designating the waters as a Marine Protected Area to safeguard the existing population and provide benefit to the surrounding marine environment.

Geographical Impact and Ecological Restoration Modes of the Spatial Differentiation of Rural Social-Ecosystem Vulnerability: Evidence from Qingpu District in Shanghai

Vulnerability research is the core issue and one of the research hotspots of sustainable development science. Vulnerability and its evaluation framework provide a new perspective for rural social-ecosystem studies. This paper introduced the ‘input-output’ efficiency theory and constructed the ‘SEE-PSR’ framework for the analysis of social-ecosystem vulnerability in the rural area in Qingpu District of Shanghai City. The DEA models, spatial autocorrelation model, multivariate logistic regression model, geographical detector and hierarchical cluster model were used to analyze the spatial differences of social-ecosystem vulnerability, and its geographical impact mechanisms and ecological restorations, in 184 administrative villages in this area. The results can be divided into three main points. (1) The results of the ‘input-output’ efficiency model of the EW-DEA based on entropy weight aggregation crossover was more reliable and accurate for the evaluation of rural social-ecosystem vulnerability. The vulnerability of the social-ecosystems in the administrative villages showed a trend of gradual decline from east to west, with an average value of vulnerability of 0.583, and the vulnerability of social systems had become an important factor in constraining the decrease of the vulnerability of the social-ecosystems in the region. (2) The distances from the center of Shanghai City, from Dianshan Lake, from the center of Qingpu District and from the water area were the four dominant geographical factors affecting the vulnerability of the social-ecosystem in this region. The geographical impacts exhibited the spatial differentiations of systemic structure, the substitution of typological attributes and the transformation level. (3) The geographical factors coupling the impact types of the social-ecosystem vulnerability were divided spatially into 10 types. The geographic multi-factor coupling impact types were dominant, which presented multi-cyclic spatial patterns and were dominated by the central multi-factor which was surrounded by the single factor types on both sides. According to the different types, some feasible ways of ecological restoration were proposed, which drew on the experiences of integrated territory consolidation to remediate the vulnerability of rural social-ecological systems. The results of this study can provide scientific reference for rural spatial reconstruction, regional ecological restoration and sustainable development for the regions characterized by conflict in the ‘strict protection of the ecological environment and vigorous development of the economy’.

Self-supervised delineation of geologic structures using orthogonal latent space projection

We have developed two machine-learning frameworks that could assist in automated lithostratigraphic interpretation of seismic volumes without any manual hand labeling from an experienced seismic interpreter. The first framework is an unsupervised hierarchical clustering model used to divide seismic images from a volume into a certain number of clusters determined by the algorithm. The clustering framework uses a combination of density and hierarchical techniques to determine the size and homogeneity of the clusters. The second framework consists of a self-supervised deep learning framework to label regions of geologic interest in seismic images. It projects the latent space of an encoder-decoder architecture onto two orthogonal subspaces, from which it learns to delineate regions of interest in the seismic images. To demonstrate an application of both frameworks, a seismic volume is clustered into various contiguous clusters, from which four clusters are selected based on distinct seismic patterns: horizons, faults, salt domes, and chaotic structures. Images from the selected clusters are used to train the encoder-decoder network. The output of the encoder-decoder network is a probability map of the possibility that an amplitude reflection event belongs to an interesting geologic structure. The structures are delineated using the probability map. The delineated images are further used to posttrain a segmentation model to extend our results to full vertical sections. The results on vertical sections indicate that we can factorize a seismic volume into its corresponding structural components. Finally, we find that our deep learning framework can be modeled as an attribute extractor and we compare our attribute result with various existing attributes in the literature and determine the competitive performance with them.

Segmentation of Natural Images with K-Means and Hierarchical Algorithm based on Mixture of Pearson Distributions

In this paper, an attempt has been made to analyze the performance of the image segmented algorithms with the addition of the Pearsonian Type III mixture model. By using the Type III Pearsonian system of distributions the image segmentation process was carried out in the current article which is a novel technique. With the help of K-component combination of Pearsonian Type III distribution, it is considered that the whole input images are characterized. The performance parameters PRI (Probabilistic Rand Index), GCE (Global Consistency Error) and VOI (Volume of Interest) for the currently considered model are estimated with the help of EM (Expectation Maximization) algorithm. For analyzing the proposed model’s performance, four random images are selected as input for the current model from Berkeley image database. The performance metric parameters PRI, GCE and VOI values given the results as the currently proposed method is providing more précise results for the input images where the regions of the input images selected are with tiles having long upper model and the left skewed images. By the help of image quality measures, the proposed method is performing well for the purpose of retrieving the images with respect to the picture segmenting process which is based on GMM (Gaussian Mixture Model). The current model performance was compared with the other existing models like the k-means hierarchical clustering model and the 3-paprameter regression models.

Teasing Apart Silvopasture System Components Using Machine Learning for Optimization

Silvopasture systems combine tree and livestock production to minimize market risk and enhance ecological services. Our objective was to explore and develop a method for identifying driving factors linked to productivity in a silvopastoral system using machine learning. A multi-variable approach was used to detect factors that affect system-level output (i.e., plant production (tree and forage), soil factors, and animal response based on grazing preference). Variables from a three-year (2017–2019) grazing study, including forage, tree, soil, and terrain attribute parameters, were analyzed. Hierarchical variable clustering and random forest model selected 10 important variables for each of four major clusters. A stepwise multiple linear regression and regression tree approach was used to predict cattle grazing hours per animal unit (h ha−1 AU−1) using 40 variables (10 per cluster) selected from 130 total variables. Overall, the variable ranking method selected more weighted variables for systems-level analysis. The regression tree performed better than stepwise linear regression for interpreting factor-level effects on animal grazing preference. Cattle were more likely to graze forage on soils with Cd levels &lt;0.04 mg kg−1 (126% greater grazing hours per AU), soil Cr &lt;0.098 mg kg−1 (108%), and a SAGA wetness index of &lt;2.7 (57%). Cattle also preferred grazing (88%) native grasses compared to orchardgrass (Dactylis glomerata L.). The result shows water flow within the landscape position (wetness index), and associated metals distribution may be used as an indicator of animal grazing preference. Overall, soil nutrient distribution patterns drove grazing response, although animal grazing preference was also influenced by aboveground (forage and tree), soil, and landscape attributes. Machine learning approaches helped explain pasture use and overall drivers of grazing preference in a multifunctional system.

Grouping of dynamic electricity consumption behaviour: An f ‐divergence based hierarchical clustering model

Under the digitalization trend in the energy sector, utilities are devoted to providing better service to their customers by mining knowledge in fine-grained electricity consumption data. Understanding the group behaviour of customers by clustering method is essential to achieving this end. Different from shape-based clustering methods, an f-divergence based hierarchical clustering model is proposed to group customers by their dynamic electricity consumption patterns. Modelling the electricity consumption by Markov chains, the customers’ consumption patterns are first summarized into transition probability matrixes. Then, dissimilarity measures based on f-divergence are calculated. Specifically, due to their superiority, squared Hellinger distance and total variation distance are used. The hierarchical clustering is then conducted based on the obtained distance matrixes. Using real-world smart meter dataset, the proposed method is compared with other dynamic clustering candidates by using the revised silhouette score. And consumers’ dynamic consumption patterns are not only analysed from the global to local levels, but also the relationship between clustering results and external factors are delved into. The results show that the proposed method can produce highly representative clusters, and is able to provide insights on the implementation of the demand-side management program.

Application of agglomerative hierarchical clustering and logistic model development for assessing solar energy acceptability as an alternate energy option

This paper is focused on the assessment of acceptability of solar energy as an alternate efficient energy management option using Agglomerative Hierarchy Cluster (AHC) and logistic regression modelling approach. The study population includes randomly selected shop-owners and residential occupants within the Port Harcourt city in Rivers State, Nigeria. The collected data sets were subjected to AHC analysis using a statistical package XLSTAT 2016 version 4.6. The central object identified from the application of AHC with respect to the sampled shop-owners and residential occupants as pertaining to the acceptability of solar energy as an alternate efficient energy management option was centered around the financial implication of energy generation and the political influence of the government solar energy policies for energy generation. Finally, logistic regression modelling approach was applied into developing a predictive model for the probability of general acceptance (variable ‘yes’) of solar energy as an effective energy management system. From the developed model the chance of acceptance of a solar energy management system is 1% with 59.5% rejection from the study population while it is 99% with an unawareness level of 40.51% from the study population.

Application of agglomerative hierarchical clustering and logistic model development for assessing solar energy acceptability as an alternate energy option

Application of agglomerative hierarchical clustering and logistic model development for assessing solar energy acceptability as an alternate energy option

Resonance as an Applied Predictor of Cross-Cultural Interaction: Constructional Priming in Mandarin and American English Interaction

AbstractIn Dialogic syntax (cf. Du Bois 2014; Tantucci et al. 2018), naturalistic interaction is inherently grounded in resonance, viz. the catalytic activation of affinities across turns (Du Bois and Giora 2014). Resonance occurs dynamically when interlocutors creatively coconstruct utterances that are formally and phonetically similar to the utterance of a prior speaker. In this study, we argue that such similarity can inform the machine learning prediction of linguistic and cross-cultural diversity. We compared two sets of 1,000 exchanges involving (dis)-agreement from the two balanced Callhome corpora of naturalistic interaction in Mandarin Chinese and American English. We found a correlation of overt use of pragmatic markers with resonance, indicating that priming does not occur as an exclusively implicit mechanism (as it is commonly held in the experimental literature e.g. Bock 1986; Bock et al. 2007), but naturalistically underpins dialogic engagement and cooperation among interactants. We fitted a mixed effects linear regression and a hierarchical clustering model to show that resonance occurs formally and functionally in different ways from one language to another. The applied results of this study can lead to a novel turn in AI research of conversational interfaces (McTear et al. 2016; Klopfenstein et al. 2017), as they reveal the fundamental role played cross-linguistically by resonance as a form of engagement of human-to-human interaction and the importance to address this mechanism in machine-to-human communication.

PURA: A Product-and-User Oriented Approach for Requirement Analysis From Online Reviews

Users often show different and personalized product requirements with respect to their gender, age, location, and other attributes. The requirement analysis of various users is thus vitally important for product evolution and customization. Accordingly, we propose a product-and-user oriented approach for requirement analysis, which can automatically elicit, classify, and rank product requirements from online reviews, and identify requirement differences among users. More specifically, the product features are first extracted from online reviews in a machine-learning manner. Next, a hierarchical clustering model is applied for product feature classification, and a weighted function of customer sentiment and attention is put forward for requirement ranking. Then, statistical methods are used for requirement preference identification. Finally, the effectiveness and feasibility of the proposed approach is examined on 2198 online reviews with 148 527 subsentences about new energy vehicles.

Identifying unknown Indian wolves by their distinctive howls: its potential as a non-invasive survey method

Previous studies have posited the use of acoustics-based surveys to monitor population size and estimate their density. However, decreasing the bias in population estimations, such as by using Capture–Mark–Recapture, requires the identification of individuals using supervised classification methods, especially for sparsely populated species like the wolf which may otherwise be counted repeatedly. The cryptic behaviour of Indian wolf (Canis lupus pallipes) poses serious challenges to survey efforts, and thus, there is no reliable estimate of their population despite a prominent role in the ecosystem. Like other wolves, Indian wolves produce howls that can be detected over distances of more than 6 km, making them ideal candidates for acoustic surveys. Here, we explore the use of a supervised classifier to identify unknown individuals. We trained a supervised Agglomerative Nesting hierarchical clustering (AGNES) model using 49 howls from five Indian wolves and achieved 98% individual identification accuracy. We tested our model’s predictive power using 20 novel howls from a further four individuals (test dataset) and resulted in 75% accuracy in classifying howls to individuals. The model can reduce bias in population estimations using Capture-Mark-Recapture and track individual wolves non-invasively by their howls. This has potential for studies of wolves’ territory use, pack composition, and reproductive behaviour. Our method can potentially be adapted for other species with individually distinctive vocalisations, representing an advanced tool for individual-level monitoring.

Incorporating statistical clustering methods into mortality models to improve forecasting performances

Statistical clustering is a procedure of classifying a set of objects such that objects in the same class (called cluster) are more homogeneous, with respect to some features or characteristics of objects, to each other than to those in any other classes. In this paper, we apply four clustering approaches to improving forecasting performances of the Lee–Carter and CBD models. First, each of four clustering methods (Ward’s hierarchical clustering, divisive hierarchical clustering, K-means clustering, and Gaussian mixture model clustering) is adopted to determine, based on some characteristics of mortality rates, the number and partition of age clusters from the whole study ages 25-84. Next, we forecast 10-year and 20-year mortality rates for each of the age clusters using the Lee–Carter and CBD models, respectively. Finally, numerical illustrations are given with two R packages “NbClust” and “mclust” for clustering. Mortality data for both genders of the US and the UK are obtained from the Human Mortality Database, and the MAPE (mean absolute percentage error) measure is adopted to evaluate forecasting performance. Comparisons of MAPE values are made with and without clustering, which demonstrate that all the proposed clustering methods can improve forecasting performances of the Lee–Carter and CBD models.

HCMMCNVs: hierarchical clustering mixture model of copy number variants detection using whole exome sequencing technology.

In this article, we introduce a hierarchical clustering and Gaussian mixture model with expectation-maximization (EM) algorithm for detecting copy number variants (CNVs) using whole exome sequencing (WES) data. The R shiny package 'HCMMCNVs' is also developed for processing user-provided bam files, running CNVs detection algorithm and conducting visualization. Through applying our approach to 325 cancer cell lines in 22 tumor types from Cancer Cell Line Encyclopedia (CCLE), we show that our algorithm is competitive with other existing methods and feasible in using multiple cancer cell lines for CNVs estimation. In addition, by applying our approach to WES data of 120 oral squamous cell carcinoma (OSCC) samples, our algorithm, using the tumor sample only, exhibits more power in detecting CNVs as compared with the methods using both tumors and matched normal counterparts. HCMMCNVs R shiny software is freely available at github repository https://github.com/lunching/HCMM_CNVs.and Zenodo https://doi.org/10.5281/zenodo.4593371. Supplementary data are available at Bioinformatics online.

Поиск заимствований в армянских текстах путем внутреннего стилометрического анализа

In this work we study the application of intrinsic stylometric methods to the task of plagiarism detection in Armenian texts. We use two task setups from PAN’s series of conferences on text forensics and stylometry: style change detection and style breach detection. Style change detection aims to determine whether the text is written by more than one author, while style breach detection detects the boundaries of stylistically distinct text fragments. For these tasks, we generate synthetic test sets for three genres of text: academic, literature, and news, and then use them to evaluate the effectiveness of hierarchical clustering and other relevant models from PAN conferences. We employ a standard set of character-level, lexical and readability features, and additionally perform morphological and dependency parsing of text fragments to extract syntactic features encoding author style information. The evaluation results show that the clustering-based approach fails to correctly detect style change detection in longer texts and is only marginally better for shorter texts. For style breach detection, hierarchical clustering-based approach performs better than a random baseline classifier, but the difference is not sufficient to warrant its practical use. In a complementary experiment, we show that reducing the number of features and multicollinearity in them via PCA helps to increase the precision of style breach detection methods for certain text categories.