Distribution Of Clusters Research Articles

A Three-Dimensional Hail Trajectory Clustering Technique

Abstract Recent advances in hail trajectory modeling regularly produce datasets containing millions of hail trajectories. Because hail growth within a storm cannot be entirely separated from the structure of the trajectories producing it, a method to condense the multidimensionality of the trajectory information into a discrete number of features analyzable by humans is necessary. This article presents a three-dimensional trajectory clustering technique that is designed to group trajectories that have similar updraft-relative structures and orientations. The new technique is an application of a two-dimensional method common in the data mining field. Hail trajectories (or “parent” trajectories) are partitioned into segments before they are clustered using a modified version of the density-based spatial applications with noise (DBSCAN) method. Parent trajectories with segments that are members of at least two common clusters are then grouped into parent trajectory clusters before output. This multistep method has several advantages. Hail trajectories with structural similarities along only portions of their length, e.g., sourced from different locations around the updraft before converging to a common pathway, can still be grouped. However, the physical information inherent in the full length of the trajectory is retained, unlike methods that cluster trajectory segments alone. The conversion of trajectories to an updraft-relative space also allows trajectories separated in time to be clustered. Once the final output trajectory clusters are identified, a method for calculating a representative trajectory for each cluster is proposed. Cluster distributions of hailstone and environmental characteristics at each time step in the representative trajectory can also be calculated. Significance Statement To understand how a storm produces large hail, we need to understand the paths that hailstones take in a storm when growing. We can simulate these paths using computer models. However, the millions of hailstones in a simulated storm create millions of paths, which is hard to analyze. This article describes a machine learning method that groups together hailstone paths based on how similar their three-dimensional structures look. It will let hail scientists analyze hailstone pathways in storms more easily, and therefore better understand how hail growth happens.

Small-angle X-ray scattering analysis on the in-plane cluster distribution during heating amorphous Mg85Y9Cu6 alloys

In-plane cluster arrangements of a long-period stacking ordered (LPSO) structure in MgYCu alloys were examined by reverse Monte Carlo fitting of small-angle X-ray scattering (SAXS) intensities. In situ SAXS measurements were performed at a heating rate of 10 K min−1. from the glass state to the mp. The clusters were characterized as so-called L12 clusters commonly observed in the LPSO structures. The results suggest that the average in-plane distance of the clusters decreased during heating. The distance showed two preferred average distances, in contrast to the MgYZn case where the average distance decrease monotonically during heating.

Identification of 2,4-diacetylphloroglucinol production in the genus Chromobacterium

The compound 2,4-diacetylphloroglucinol (DAPG) is a broad-spectrum antibiotic that is primarily produced by Pseudomonas spp. DAPG plays an important role in the biocontrol disease suppressing activity of Pseudomonas spp. In the current study, we report the discovery of the DAPG biosynthetic cluster in strains of Chromobacterium vaccinii isolated from Brazilian aquatic environments and the distribution of the biosynthetic cluster in the Chromobacterium genus. Phylogenetic analysis of the phlD protein suggests the biosynthetic cluster probably entered the genus of Chromobacterium after a horizontal gene transfer event with a member of the Pseudomonas fluorescens group. We were able to detect trace amounts of DAPG in wild type cultures and confirm the function of the cluster with heterologous expression in Escherichia coli. In addition, we identified and verified the presence of other secondary metabolites in these strains. We also confirmed the ability of C. vaccinii strains to produce bioactive pigment violacein and bioactive cyclic depsipeptide FR900359. Both compounds have been reported to have antimicrobial and insecticidal activities. These compounds suggest strains of C. vaccinii should be further explored for their potential as biocontrol agents.

Insight into the Characteristics and Triggers of Loess Landslides during the 2013 Heavy Rainfall Event in the Tianshui Area, China

The 2013 heavy rainfall event (from June to July) in the Tianshui area triggered the most serious rainfall-induced group-occurring landslides since 1984, causing extensive casualties and economic losses. To better understand the characteristics and triggers of these loess landslides, we conducted a detailed analysis of the landslides and relevant influencing factors. Based on the detailed rainfall-induced landslide database obtained using visual interpretation of remote sensing images before and after rainfall, the correlation between the landslide occurrence and different influencing factors such as terrain, geomorphology, geology, and rainfall condition was analyzed. This rainfall event triggered approximately 54,000 landslides with a total area of 67.9 km2, mainly consisting of shallow loess landslides with elongated type, shallow rockslides, collapses, and mudflows. The landslides exhibited a clustered distribution, with the majority concentrated in two specific areas (i.e., Niangniangba and Shetang). The abundance index of landslides was closely associated with the hillslope gradient, total rainfall, and drainage (river) density. The landslide area density (LAD) was positively correlated with these influential factors, characterized by either an exponential or a linear relationship. The Middle Devonian Shujiaba formation (D2S) was identified to be highly susceptible to landslides, and the landslide events therein accounted for 35% of the total landslide occurrences within 22% of the study area. In addition, the E-SE aspect was more prone to landslides, while the W-NW aspect exhibited a low abundance of landslides.

A Systematic Review of The Recent Geospatial Approach in Addressing Spatially-Related Radicalism And Extremism Issues

This systematic review article focuses on the geospatial issues of radicalism and extremism. The scholar has intensified the application of geospatial in radicalism and extremism study to understand better the causes, patterns, and trends of the radicalism and extremism incidents. The advanced geospatial approach provides more spatio-temporal information on radicalism and extremism incidents'. It improves the conventional study method that only focuses on fundamentals and theory. Unfortunately, some geospatial issues from previous radicalism and extremism studies have been found. Hence, the present study reviewed past studies on geospatial applications in radicalism and extremism. Meanwhile, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) method were used to review the current research. This systematic review utilises two major journal databases, Scopus and Web of Science. Searching works found in a total of 24 articles can be analysed systematically. The selected article was separated into four corresponding geospatial analysis types: distribution pattern analysis, cluster analysis, statistical and prediction analysis, and 3D technology. Finally, several recommendations were offered after this study for future scholars' consideration.

A Data Enhancement Algorithm for DDoS Attacks Using IoT.

With the rapid development of the Internet of Things (IoT), the frequency of attackers using botnets to control IoT devices in order to perform distributed denial-of-service attacks (DDoS) and other cyber attacks on the internet has significantly increased. In the actual attack process, the small percentage of attack packets in IoT leads to low accuracy of intrusion detection. Based on this problem, the paper proposes an oversampling algorithm, KG-SMOTE, based on Gaussian distribution and K-means clustering, which inserts synthetic samples through Gaussian probability distribution, extends the clustering nodes in minority class samples in the same proportion, increases the density of minority class samples, and improves the amount of minority class sample data in order to provide data support for IoT-based DDoS attack detection. Experiments show that the balanced dataset generated by this method effectively improves the intrusion detection accuracy in each category and effectively solves the data imbalance problem.

Development and testing the feasibility of a sports-based mental health promotion intervention in Nepal: a protocol for a pilot cluster-randomised controlled trial

BackgroundMental wellbeing encompasses life satisfaction, social connectedness, agency and resilience. In adolescence, mental wellbeing reduces sexual health risk behaviours, substance use and violence; improves educational outcomes; and protects mental health in adulthood. Mental health promotion seeks to improve mental wellbeing and can include activities to engage participants in sport. However, few high-quality trials of mental health promotion interventions have been conducted with adolescents, especially in low- and middle-income countries. We sought to address this gap by testing SMART (Sports-based Mental heAlth pRomotion for adolescenTs) in a pilot cluster-randomised controlled trial (cRCT) in Bardiya, Nepal.MethodsThe objectives of the trial are to assess the acceptability and feasibility of SMART, test trial procedures, explore outcome distributions in intervention and control clusters and calculate the total annual cost of the intervention and unit cost per adolescent. The trial design is a parallel-group, two-arm superiority pilot cRCT with a 1:1 allocation ratio and two cross-sectional census surveys with adolescents aged 12–19, one pre-intervention (baseline) and one post-intervention (endline). The study area is four communities of approximately 1000 population (166 adolescents per community). Each community represents one cluster. SMART comprises twice weekly football, martial arts and dance coaching, open to all adolescents in the community, led by local sports coaches who have received psychosocial training. Sports melas (festivals) and theatre performances will raise community awareness about SMART, mental health and the benefits of sport. Adolescents in control clusters will participate in sport as usual. In baseline and endline surveys, we will measure mental wellbeing, self-esteem, self-efficacy, emotion regulation, social support, depression, anxiety and functional impairment. Using observation checklists, unstructured observation and attendance registers from coaching sessions, and minutes of meetings between coaches and supervisors, we will assess intervention fidelity, exposure and reach. In focus group discussions and interviews with coaches, teachers, caregivers and adolescents, we will explore intervention acceptability and mechanisms of change. Intervention costs will be captured from monthly project accounts, timesheets and discussions with staff members.DiscussionFindings will identify elements of the intervention and trial procedures requiring revision prior to a full cRCT to evaluate the effectiveness of SMART.Trial registrationISRCTN, ISRCTN15973986, registered on 6 September 2022; ClinicalTrials.gov, NCT05394311, registered 27 May 2022.

Quantifying the spatial clustering characteristics of radiographic emphysema explains variability in pulmonary function

Quantitative assessment of emphysema in CT scans has mostly focused on calculating the percentage of lung tissue that is deemed abnormal based on a density thresholding strategy. However, this overall measure of disease burden discards virtually all the spatial information encoded in the scan that is implicitly utilized in a visual assessment. This simplification is likely grouping heterogenous disease patterns and is potentially obscuring clinical phenotypes and variable disease outcomes. To overcome this, several methods that attempt to quantify heterogeneity in emphysema distribution have been proposed. Here, we compare three of those: one based on estimating a power law for the size distribution of contiguous emphysema clusters, a second that looks at the number of emphysema-to-emphysema voxel adjacencies, and a third that applies a parametric spatial point process model to the emphysema voxel locations. This was done using data from 587 individuals from Phase 1 of COPDGene that had an inspiratory CT scan and plasma protein abundance measurements. The associations between these imaging metrics and visual assessment with clinical measures (FEV_1, FEV_1-FVC ratio, etc.) and plasma protein biomarker levels were evaluated using a variety of regression models. Our results showed that a selection of spatial measures had the ability to discern heterogeneous patterns among CTs that had similar emphysema burdens. The most informative quantitative measure, average cluster size from the point process model, showed much stronger associations with nearly every clinical outcome examined than existing CT-derived emphysema metrics and visual assessment. Moreover, approximately 75% more plasma biomarkers were found to be associated with an emphysema heterogeneity phenotype when accounting for spatial clustering measures than when they were excluded.

Trust based attachment.

In social systems subject to indirect reciprocity, a positive reputation is key for increasing one's likelihood of future positive interactions [1-13]. The flow of gossip can amplify the impact of a person's actions on their reputation depending on how widely it spreads across the social network, which leads to a percolation problem [14]. To quantify this notion, we calculate the expected number of individuals, the "audience", who find out about a particular interaction. For a potential donor, a larger audience constitutes higher reputational stakes, and thus a higher incentive, to perform "good" actions in line with current social norms [7, 15]. For a receiver, a larger audience therefore increases the trust that the partner will be cooperative. This idea can be used for an algorithm that generates social networks, which we call trust based attachment (TBA). TBA produces graphs that share crucial quantitative properties with real-world networks, such as high clustering, small-world behavior, and powerlaw degree distributions [16-21]. We also show that TBA can be approximated by simple friend-of-friend routines based on triadic closure, which are known to be highly effective at generating realistic social network structures [19, 22-25]. Therefore, our work provides a new justification for triadic closure in social contexts based on notions of trust, gossip, and social information spread. These factors are thus identified as potential significant influences on how humans form social ties.

Relationship chains of subhealth physical examination indicators: a cross-sectional study using the PLS-SEM approach

Subhealth is a transitional state between health and disease, and it can be detected through routine physical check-ups. However, the complexity and diversity of physical examination items and the difficulty of quantifying subhealth manifestations are the main problems that hinder its treatment. The aim of this study was to systematically investigate the physical examination performance of the subhealthy population and further explore the deeper relationships between indicators. Indicators were obtained for 878 subjects, including basic information, Western medicine indicators, inquiries of traditional Chinese medicine and sublingual vein (SV) characteristics. Statistical differences were analysed using R software. To explore the distribution of symptoms and symptom clusters in subhealth, partial least squares-structural equation modelling (PLS-SEM) was applied to the subhealth physical examination index, and a structural model was developed to verify whether the relationship chain between the latent variables was reasonable. Finally, the reliability and validity of the PLS-SE model were assessed. The most common subclinical clinical symptoms were limb soreness (37.6%), fatigue (31.6%), shoulder and neck pain (30.5%) and dry eyes (29.2%). The redness of the SV in the subhealthy group was paler than that in the healthy group (p < 0.001). This study validates the establishment of the directed acyclic relationship chain in the subhealthy group: the path from routine blood tests to lipid metabolism (t = 7.878, p < 0.001), the path from lipid metabolism to obesity (t = 8.410, p < 0.001), the path from obesity to SV characteristics (t = 2.237, p = 0.025), and the path from liver function to SV characteristics (t = 2.215, p = 0.027). The innovative application of PLS-SEM to the study of subhealth has revealed the existence of a chain of relationships between physical examination indicators, which will provide a basis for further exploration of subhealth mechanisms and causal inference. This study has identified the typical symptoms of subhealth, and their early management will help to advance the treatment of diseases.

Association between mental health workforce supply and clusters of high and low rates of youth suicide: An Australian study using suicide mortality data from 2016 to 2020.

To examine the association between mental health workforce supply and spatial clusters of high versus low incidence of youth suicide. A cross-sectional analysis of spatial suicide clusters in young Australians (aged 10-25) from 2016 to 2020 was conducted using the scan statistic and suicide data from the National Coronial Information System. Mental health workforce was extracted from the 2020 National Health Workforce Dataset by local government areas. The Geographic Index of Relative Supply was used to estimate low and moderate-to-high mental health workforce supply for clusters characterised by a high and low incidence of suicide (termed suicide hotspots and coldspots, respectively). Univariate and multivariate logistic regression was used to determine the association between suicide clusters and a range of sociodemographic characteristics including mental health workforce supply. Eight suicide hotspots and two suicide coldspots were identified. The multivariate analysis showed low mental health workforce supply was associated with increased odds of being involved in a suicide hotspot (adjusted odds ratio = 8.29; 95% confidence interval = 5.20-13.60), followed by residential remoteness (adjusted odds ratio = 2.85; 95% confidence interval = 1.68-4.89), and illicit drug consumption (adjusted odds ratio = 1.97; 1.24-3.11). Both coldspot clusters occurred in areas with moderate-to-high mental health workforce supply. Findings highlight the potential risk and protective roles that mental health workforce supply may play in the spatial distributions of youth suicide clusters. These findings have important implications for the provision of postvention and the prevention of suicide clusters.

Spatio-temporal analysis of the impact of land urbanization on the gross primary productivity of vegetation in the Middle Reaches of the Yangtze River Urban Agglomeration: new evidence from the township scale

The urgent need to maintain ecosystem provisioning services and achieve urban sustainable development goals has led to a long-standing focus on the driving effects of land urbanization (LU) in the Middle Reaches of the Yangtze River Urban Agglomeration (MRYRU) on vegetation cover system productivity. Previous studies have lacked consideration for the long-term correlation between land urbanization (LU) and the Gross Primary Productivity (GPP) of vegetation, as well as the spatial non-stationarity of LU’s impact in the region. This study is based on land cover remote sensing data and GPP imagery data to monitor the long-term evolution trends of LU and GPP in MRYRU from the perspectives of townships and grids. It further investigates the spatial correlation and clustering characteristics between the two using bivariate spatial autocorrelation method. Additionally, a Geographically and Temporally Weighted Regression (GTWR) model was employed to analyze the spatial effects of LU on GPP. The results indicate that the GPP of MRYRU grows to 1572.88gCm-2a-1 with an average annual growth rate of 0.848% from 2000 to 2020. High-value areas are distributed in the central and eastern mountainous of the urban agglomeration, while low-value areas gradually formed an “n”-shaped clustered distribution pattern. The negative spatial autocorrelation between LU and GPP shows a general upward trend, and Low-High is the dominant agglomeration type, concentrated in Hengyang and some mountainous cities in Jiangxi Province. The overall negative impact of land urbanization on GPP in MRYRU is maintained in the spatially nonstationary pattern of driving coefficients, with negative impact areas mostly concentrated in the city’s main city zone and positive impact areas concentrated in the urban agglomeration’s mountainous townships. This study provides support for MRYRU to adopt more resilient land optimization management strategies and vegetation ecological restoration plans.

Accounting for diverse feature-types improves patient stratification on tabular clinical datasets

Tabular Clinical and Biomedical Routine Data (CBRD) contains diverse feature types. Recent research shows that the conventional application of Uniform Manifold Projection and Approximation (UMAP) to extract clusters from the low dimensional embedding can prove ineffective due to the diverse feature types in such datasets. Feature-type Distributed Clustering (FDC) workflow accounts for these diverse feature types resulting in a more informative low-dimensional embedding. However, a rigorous assessment of the FDC algorithm is missing so far. In this work, we conducted comprehensive benchmarking experiments to compare the quality of the cluster distributions and low dimensional embeddings generated by the FDC against that of the ones generated by UMAP using standard objective measures: Silhouette score, Dunn index, and ANOVA. Our results confirm that FDC can indeed be the better choice to embed tabular data with diverse feature types in low dimensions and thereby extract clusters from such an embedding. In addition, we provide a rationale behind the choice of metrics proposed in the FDC workflow. Moreover, we also point out some problems with the original Canberra metric used to reduce ordinal features in the FDC workflow and provide a solution in the form of a modified version of the Canberra metric. Using seven datasets from the medical domain for benchmarking, we demonstrate that FDC leads to improved patient stratification.

Observing Etna volcano dynamics through seismic and deformation patterns

Geophysical data provide the chance to investigate a volcano’s dynamics; considerable information can especially be gleaned on the stress and strain patterns accompanying the internal processes and the effect of magma ascent on the main structures triggering earthquakes. Here, we analysed in detail the seismicity recorded over the last two decades on Etna volcano (southern Italy), focusing on earthquakes distribution and focal mechanism clustering; the ground deformation pattern affecting the volcanic edifice with the inflation and deflation phases was also examined. Analysed data were compared in order to shed light on possible relationships with the volcanic activity and to better understand the internal dynamics of the volcano over time. Significant steps during or shortly before major eruptions in the seismic strain release and ground deformation temporal series highlight a straightforward relationship between seismicity occurring at shallow level, inflation/deflation and volcanism. Furthermore, at depths greater than 5–7 km, down to about 20 km, the orientation of the P- and T-axes clearly indicate the existence of a pressure source in the central part of the volcano. All the results underline that the stress field related to the volcano plumbing system interferes with the regional field, partly overriding it.

Spatial distribution and biological characteristics of the Okhotsk Sea Pollock in the winter-spring period

The data on the biology and fishing of pollock in the northern part of the Sea of Okhotsk collected during its monitoring on fishing vessels in the winter-spring period 2001–2021 are analyzed. Analysis of the dislocation of the fishing fleet makes it possible to trace the formation and distribution of pollock clusters. A significant amount of the collected material made it possible to characterize the dynamics of gonad maturation and the change in the size and age structure of individuals in commercial subzones. Pollock fishing usually began in the Kamchatka-Kuril subzone and partly in the North Okhotsk subzone in the area of the Swan Upland, with gradual development northward into Shelikhov Bay, Pritaui district and the North Okhotsk shelf. The basis of the catches was pollock at the age of 5–8 years at the pre-spawning stages of the development of sexual products. By-catch of juveniles varied from 0,5% (2009) to 63,5% (2018), averaging 23,5% in the North Okhotsk subzone, 34,8% in the West Kamchatka subzone and 15,5% in the Kamchatka-Kuril subzone.

Analysis of the spatial-temporal distribution characteristics of hepatitis E in Jiangsu province from 2005 to 2020.

This study attempts to analyze the spatial clustering and spatial-temporal distribution characteristics of hepatitis E (HE) at the county (city and district) level in Jiangsu province to provide a scientific basis for the prevention and control of HE. The information on HE cases reported in the Chinese Center for Disease Control and Prevention Information System from 2005 to 2020 was collected for spatial autocorrelation analysis and spatial-temporal clustering analysis. From 2005 to 2020, 48,456 HE cases were reported in Jiangsu province, with an average annual incidence rate of 3.87/100,000. Male cases outnumbered female cases (2.46:1), and the incidence was highest in the 30-70 years of age group (80.50%). Farmers accounted for more than half of all cases (59.86%), and in terms of the average annual incidence, the top three cities were all in Zhenjiang city. Spatial autocorrelation analysis showed that Global Moran's I of HE incidence varied from 0.232 to 0.513 for the years. From 2005 to 2020, 31 counties (cities and districts) had high and statistically significant HE incidence, and two clustering areas were detected by spatial-temporal scanning. HE incidence in Jiangsu province from 2005 to 2020 was stable, with age and gender differences, regional clustering, and spatial-temporal clustering. Further investigation of HE clustering areas is necessary to formulate corresponding targeted prevention and control measures.

Statistical Analysis of Tribological Properties of Mg(AM50)/GNF-Al2O3sf Hybrid Composites

The present article describes the tribological properties of Mg-based hybrid composites reinforced with graphite nanofiber (GNF) and alumina short fiber (Al2O3sf) that were investigated. The Mg/GNF/Al2O3sf hybrid composites with varying volume fraction of fiber (10 vol.%, 15 vol.%, 20 vol.%) were developed. SEM observations indicate that the GNF cluster distributions within the array of the Al2O3sf network are found to be relatively good. The Taguchi design of the experiment has been applied to conduct the wear test, and the statistical analysis of variance (ANOVA) has been used to evaluate the influence of wear test parameters on the wear loss and coefficient of friction (COF) of the composites. The influence of wear test parameters such as volume fraction of fiber (VF), applied load (AL), sliding distance (SD), and sliding speed (SP) on the wear loss and COF of composites was analyzed under dry sliding conditions. The results of ANOVA indicate that the sliding distance was found to be the prominent factor affecting wear loss, and the applied load influenced the COF most significantly. Furthermore, the composites with 20 vol.% of fiber had lower wear loss than those with 10 vol.% and 15 vol.% of fiber. The COF of composites with 15 vol.% of fiber was found to be slightly lower compared to the 10 vol.% and 20 vol.% of fiber cases. The results imply that the hybridization of GNFs and Al2O3sf, as well as the formation of Mg17Al12 and Al2MgC2 precipitates enhanced the tribological properties of the Mg hybrid composites.

Analysis of U-V2X Communications with Non-Clustered and Clustered Jamming in the Presence of Fluctuating UAV Beam Width

Jammers emit strong intentional jamming signals aiming to limit or block legitimate communications. The distribution of jammers, whether in non-clustered or clustered form, significantly influences the performance of vehicle-to-everything (V2X) networks. In addition, the fluctuations in the three-dimensional (3D) antenna beam width of unmanned aerial vehicles (UAVs) can exert a substantial impact on the network’s overall performance. This paper introduces a model for UAV-V2X (U-V2X) communications in mm-Wave bands, considering non-clustered and clustered jammers, as well as the varying 3D antenna beam width. The roads are modeled using a Poisson line process, vehicular nodes (VNs) are modeled using a 1D Poisson point process (PPP), and UAVs are modeled using a 3D PPP. The jammers are distributed in two ways: non-clustered and clustered distributions. Moreover, the fluctuations in the 3D antenna beam width follow a normal distribution. To this end, a typical node’s performance in U-V2X communications is evaluated for various network configurations, including the number of UAVs, VNs, roads, jammers, and jammer’s transmission power. The analytical expressions for the outage probability (OP) of VN to VN connection (i.e., V2V), VN to UAV connection (i.e., V2U2V), and an overall connection (i.e., U-V2X), under non-clustered and clustered jamming, along with the fluctuating antenna beam width, are derived. The results revealed that the performance of the U-V2X communications utilizing mm-Waves is significantly degraded with the non-clustered jamming in comparison with the clustered jamming. The fluctuations in the 3D beam width of the UAV antennas further compromise the network’s performance. Thus, accurate modeling of these fluctuations is crucial, particularly in the presence of non-clustered jammers. Furthermore, the system designers should focus on implementing additional anti-jamming countermeasures specifically targeting non-clustered jammers in U-V2X communications.

Genomics of Acinetobacter baumannii iron uptake.

Iron is essential for growth in most bacteria due to its redox activity and its role in essential metabolic reactions; it is a cofactor for many bacterial enzymes. The bacterium Acinetobacter baumannii is a multidrug-resistant nosocomial pathogen. A. baumannii responds to low iron availability imposed by the host through the exploitation of multiple iron-acquisition strategies, which are likely to deliver iron to the cell under a variety of environmental conditions, including human and animal infection. To date, six different gene clusters for active iron uptake have been described in A. baumannii , encoding protein systems involved in (i) ferrous iron uptake (feo); (ii) haem uptake (hemT and hemO); and (iii) synthesis and transport of the baumannoferrin(s) (bfn), acinetobactin (bas/bau) and fimsbactin(s) (fbs) siderophores. Here we describe the structure, distribution and phylogeny of iron-uptake gene clusters among &gt;1000 genotypically diverse A. baumannii isolates, showing that feo, hemT, bfn and bas/bau clusters are very prevalent across the dataset, whereas the additional haem-uptake system hemO is only present in a portion of the dataset and the fbs gene cluster is very rare. Since the expression of multiple iron-uptake clusters can be linked to virulence, the presence of the additional haem-uptake system hemO may have contributed to the success of some A. baumannii clones.

Entropy regularization in probabilistic clustering

Bayesian nonparametric mixture models are widely used to cluster observations. However, one major drawback of the approach is that the estimated partition often presents unbalanced clusters’ frequencies with only a few dominating clusters and a large number of sparsely-populated ones. This feature translates into results that are often uninterpretable unless we accept to ignore a relevant number of observations and clusters. Interpreting the posterior distribution as penalized likelihood, we show how the unbalance can be explained as a direct consequence of the cost functions involved in estimating the partition. In light of our findings, we propose a novel Bayesian estimator of the clustering configuration. The proposed estimator is equivalent to a post-processing procedure that reduces the number of sparsely-populated clusters and enhances interpretability. The procedure takes the form of entropy-regularization of the Bayesian estimate. While being computationally convenient with respect to alternative strategies, it is also theoretically justified as a correction to the Bayesian loss function used for point estimation and, as such, can be applied to any posterior distribution of clusters, regardless of the specific model used.