Class Co-occurrence Research Articles

Large-scale analysis of the PAC domain structure of arogenate dehydratases reveals their evolutionary patterns in angiosperms

Arogenate dehydratase (ADT) is the key limiting enzyme of plant phenylalanine biosynthesis, but some ADTs display a prephenate decarboxylase/dehydratase activity-conferring (PAC) domain. The genome resources of 70 species were employed to identify genes and outline their characteristics, especially the number and type of PAC domain structures. We obtained 522 ADTs, and their size, exon number, amino acid number and putative protein isoelectric point greatly varied from 306 to 2520 bp, 1 to 15, 101 to 839 and 4.37 to 11.18, respectively. We classified the ADTs into Class α (without a PAC domain) (115, 22.0 %), β (with a type I PAC domain) (244, 46.7 %) and γ (with a type II PAC domain) (163, 31.2 %), and their distribution frequencies exhibited large differences among various branches of angiosperms. We found that Class γ members are more conserved than Class β members, although they commonly experienced multiple duplication events and strong purifying selection, which resulted in a small number, and the putative origin order was from Class α to β and then to γ. In addition, the co-occurrence of both Class β and γ members could ensure the survival of angiosperms, while their optimized composition and strategically intertwined regulation may facilitate core eudicot success.

Ecological co-occurrence and soil physicochemical factors drive the archaeal community in Amazonian soils.

We evaluated the co-occurrence of archaeal taxonomic groups and soil physicochemical characteristics in relation to the structuring of the archaeal community in Amazonian soil under different land use systems. Soil samples were collected in primary forest (PF), secondary forest (SF), agricultural systems (AG) and cattle pastures (PA). Archaeal community composition was revealed based on high-throughput amplicon sequencing of the 16S rRNA gene. The results revealed co-occurrence of archaeal classes, with two groups formed: Thaumarchaeota classes, including South Africa Gold Mine-Group 1 (SAGMG-1), Crenarchaeotic group (SCG) and Crenarchaeota candidate division YNPFFA, with predominance in PF and SF; and Bathyarchaeota_unclassified, Methanomicrobia and Methanobacteria (Euryarchaeota) with the FHMa11 terrestrial group, with predominance in PA. The number of co-occurrences between groups was lower in SF, AG and PA (approximately 30%) than in PF. The qPCR analysis revealed that PF also had the largest number of archaeal representatives. Soil texture may be a limiting factor of interactions between groups since the most representative groups, SAGMG-1 and the SCG (over 20% in all sites), were positively associated with coarse sand, the soil factor most correlated with the groups (33% of the total). These results suggest that interactions between archaeal classes belonging to different phyla may be dependent on the number of individuals in the soil environment. In this context, differences in soil physical structure among the land use systems can reduce the representatives of key groups and consequently the co-occurrence of Archaea, which could compromise the natural dynamics of this complex environment.

Pollution and edaphic factors shape bacterial community structure and functionality in historically contaminated soils

Studies about biodegradation potential in soils often refer to artificially contaminated and simplified systems, overlooking the complexity associated with contaminated sites in a real context. This work aims to provide a holistic view on microbiome assembly and functional diversity in the model site SIN Brescia-Caffaro (Italy), characterized by historical and uneven contamination by organic and inorganic compounds. Here, physical and chemical analyses and microbiota characterization were applied on one-hundred-twenty-seven soil samples to unravel the environmental factors driving bacterial community assembly and biodegradation potential in three former agricultural fields. Chemical analyses showed a patchy distribution of metals, metalloids and polychlorinated biphenyls (PCB) and allowed soil categorization according to depth and area of collections. Likewise, the bacterial community structure, described by molecular fingerprinting and 16S rRNA gene analyses, was significantly different according to collection site and depth. Pollutant concentrations (i.e., hexachloro-biphenyls, arsenic and mercury), nitrogen content and parameters related to soil texture were identified as main drivers of microbiota assembly, being significantly correlated to bacterial community composition. Moreover, bacteria putatively involved in the aerobic degradation of PCBs were enriched over the total bacterial community in topsoils, where the highest activity was recorded using fluorescein hydrolysis as proxy. Metataxonomic analyses revealed the presence of bacteria having metabolic pathways related to PCB degradation and tolerance to heavy metals and metalloids in the topsoil samples collected in all areas. Overall, the provided dissection of soil microbiota structure and its degradation potential in the SIN Brescia-Caffaro can contribute to target specific areas for rhizoremediation implementation. Metagenomics studies could be implemented in the future to understand if specific degradative pathways are present in historically polluted sites characterized by the co-occurrence of multiple classes of contaminants.

Co-production of Classes A and B Carbapenemases BKC-1 and VIM-2 in a Clinical Pseudomonas Putida Group Isolate from Brazil.

Emergence of resistance to classical antimicrobial agents is a public health issue, especially in countries with high antimicrobial consumption rates. Carbapenems have been employed as first-choice option for empirical treatment complicated infections. However, in the last decades, frequency of carbapenemase-producing Gram-negative bacteria has rising, demanding the use of alternative antimicrobial agents. By sequencing the entire genomes with short and long reads technologies, we report the isolation and genomic characterization of a carbapenem-resistant Pseudomonas clinical isolate. The identification based on average nucleotide identity indicates a putative new species into the Pseudomonas putida Group, which carries both the blaBKC-1 and blaVIM-2 carbapenemase genes. The blaBKC-1 was found to be on a transferable IncQ plasmid backbone, whereas blaVIM-2 was found in a new integron, In2126 (intl1∆-blaVIM-2-aacA7-blaVIM-2∆-aacA27-3'CS), described in this study. Our findings indicate that co-occurrence of classes A and B carbapenemase enzymes underscores the evolving emergence of more complex antimicrobial resistance in opportunistic pathogens.

The line of fifths and the co-evolution of tonal pitch-classes

In this study, we determine the fundamental role of the line of fifths for the organization of tonal material by applying dimensionality reduction to a large historical corpus of pitch-class counts (ca. 1360–1940). We observe a historically growing trend in the exploitation of the fifths range, i.e. the size of segments that pitch-class distributions cover on the line of fifths. Moreover, we introduce the novel concept of pitch-class (co-)evolution, which traces the changing co-occurrence of pitch classes over time and likewise reaffirms the centrality of this linear tonal space from a historical angle, allowing us also to distinguish between historical periods in terms of the usage of pitch classes.

Incidence and Molecular Characterization of Carbapenemase Genes in Association with Multidrug-Resistant Clinical Isolates of Pseudomonas aeruginosa from Tertiary Healthcare Facilities in Southwest Nigeria.

Pseudomonas aeruginosa, resistant to multiple antibacterial agents including carbapenems, is of great global public health concern. There is limited data available regarding incidence of Metallo-Beta Lactamase producing P. aeruginosa, their molecular basis of resistance in particular carbapenem resistance and any genetic relatedness among circulating clinical isolates in Southwest Nigeria. Four hundred and thirty P. aeruginosa isolates were collected from seven tertiary care hospitals (predominantly from wound, ear, and urinary tract infections) and verified by PCR targeting oprI and oprL. Antibiotic susceptibility using 16 selected antibiotics and MBL screening was performed. The integrons (class 1, 2 and 3) and carbapenemase genes- blaGES, blaNMC-A, blaBIC-1, blaSME, blaIMP, blaVIM, blaSPM, blaNDM, blaAIM, blaDIM, blaSIM, blaGIM, blaOXA-48, blaOXA-58 were detected by PCR and were sequenced. Quantitative real-time polymerase chain reaction was used to quantify expression levels of eight efflux pump genes, ampC cephalosporinase and outer membrane porin, oprD. The isolates were genotyped using Enterobacterial Repetitive Intergenic Consensus sequence Polymerase Chain Reaction (ERIC-PCR). Four hundred and thirty P. aeruginosa isolates were subjected to antibiotic susceptibility testing, revealing that 109 (25.4%) isolates were multidrug-resistant, 47 (10.9%) were extensively drug-resistant and 25 (5.8%) were pandrug-resistant. MBL was seen in 17.0% (73/430) isolates. MBL-encoding genes; blaVIM-5 and blaNDM-1 were detected in 86.3% (63/73) isolates, with blaVIM-5 and blaNDM-1 in 35.6% (26/73) and 38.4% (28/73), respectively, whereas co-occurrence of blaVIM-5 and blaNDM-1 was found in 12.3% (9/73). Forty-one (56.2%) carbapenem-resistant P. aeruginosa strains carried class 1 integrons, while co-occurrence of class 1 and 2 integrons was seen in 12.3%. qPCR results indicated that MexXY-OprM was highly expressed pump in 58.9%, ampC upregulated in 26.0%, while oprD porin was downregulated in 65.8% isolates. ERIC-PCR results suggest that carbapenem-resistant strains exhibit genetic heterogeneity. The high incidence of MBL-encoding genes and integrons in diversified clinical P. aeruginosa from southwestern Nigeria is of great concern. The co-occurrence of blaVIM-5 and blaNDM-1 as well as resistance in general manifesting a gradient based on genotypic variation suggests that there is a strong need for efficient surveillance programs and antibiotic stewardship.

Leveraging network using controlled weight learning approach for thyroid cancer lymph node detection

Identification of cervical metastatic lymph nodes (LN) on I-131 post-ablation whole-body planar scans (WBS) for cancer staging is crucial for patients with papillary thyroid cancer (PTC). The existing deep network is plagued by instability in finding the under-represented LN classes on highly complex WBS, where end-to-end tuned models that can account for this uncertainty region for multi-class locations are needed. Hence, as a key contribution of this study, we designed a novel leveraging segmentation network with input guidance (LSIG) end-to-end training model without pre and post-processing features that can learn ideal parameter settings depending on the quantity of multiple-object instances. To improve the co-occurrence of classes and control the false positive regions, we proposed a re-weighting negative control (RNC) mechanism that combines two key components, namely the re-weighting (Rw) term and the negative control function (NcF). This unified approximation of weighted training would leverage the network to control and learn the desired weights of true positives towards the LN region. As an end-to-end network training, we utilize a Unet-like convolution neural network (CNN) model. The performance of the LSIG is compared with the CNN-based networks, based on the ground truth (GT) mask developed using post-ablation single-photon emission computed tomography (SPECT/CT). Furthermore, the effectiveness of the two components used in the LSIG framework is evaluated on WBS datasets. Our proposed LSIG with a fully guided (Fg) LS-FgCNN model yielded a superior performance with high AUC value of 94.9%, which is 14.6% higher than the previous network for PTC.

Co-occurrence subgroups of child sexual abuse, health risk behaviors and their associations among secondary school students in China

BackgroundLittle is known on the co-occurrence and heterogeneity of child sexual abuse (CSA) or health risk behavior (HRB) prevalence nor the associations among the victims.ObjectivesTo detect the prevalence and subgroups of adolescents reporting CSAs or HRBs, and to examine the association between the subgroups.MethodsParticipants were secondary school students in a national survey in China (N = 8746). Self-reported CSA and HRB experiences were collected through a computer assisted questionnaire. Prevalence and confidence intervals were calculated. Multigroup latent class analysis (LCA) was used to examine latent subgroups of CSA and HRB. Dual latent class regression analysis was used to examine the association between CSA and HRB classes.ResultsA total of 8746 students participated in our study. The prevalence of having ever experienced any of the reported seven CSA items was 12.9%. The preferred LCA model consisted of a three-class CSA latent variable, i.e. “Low CSAs”(95.7% of the total respondents), “Verbal or exhibitionism CSAs”(3.3%), and “high multiple CSAs” (1.1%); and a three-class HRB latent variable, i.e. “Low HRBs”(70.5%), “externalizing HRBs” (20.7%), and “internalizing HRBs” (8.7%). Students in the “Verbal or exhibitionism CSAs” or “high multiple CSAs” classes had higher probabilities of being in “externalizing HRBs” or “internalizing HRBs” classes. The probabilities were higher in “high multiple CSAs” class(male externalizing OR 4.05, 95%CI 1.71–9.57; internalizing OR 11.77, 95%CI 4.76–29.13; female externalizing OR 4.97, 95%CI 1.99–12.44; internalizing OR 9.87, 95%CI 3.71–26.25) than those in “Verbal or exhibitionism CSA”(male externalizing OR 2.51, 95%CI 1.50–4.20; internalizing OR 3.08, 95%CI 1.48–6.40; female externalizing OR 2.53, 95%CI 1.63–3.95; internalizing OR 6.05, 95%CI 3.73–9.80).ConclusionsPrevalence of CSA items varies. Non-contact CSAs are the most common forms of child sexual abuse among Chinese school students. There are different latent class co-occurrence patterns of CSA items or HRB items among the respondents. CSA experiences are in association with HRB experiences and the associations between latent classes are dose-responded. Multi-victimization has more significantly negative effects. The results could help identify high-risk subgroups and promote more nuanced interventions addressing adverse experiences and risk behaviors among at-risk adolescents.

Mental Health Risks after Repeated Exposure to Multiple Stressful Events during Ongoing Social Unrest and Pandemic in Hong Kong: The Role of Rumination: Risques pour la santé mentale après une exposition répétée à de multiples événements stressants d'agitation sociale durable et de pandémie à Hong Kong: le rôle de la rumination.

The co-occurrence of different classes of population-level stressors, such as social unrest and public health crises, is common in contemporary societies. Yet, few studies explored their combined mental health impact. The aim of this study was to examine the impact of repeated exposure to social unrest-related traumatic events (TEs), coronavirus disease 2019 (COVID-19) pandemic-related events (PEs), and stressful life events (SLEs) on post-traumatic stress disorder (PTSD) and depressive symptoms, and the potential mediating role of event-based rumination (rumination of TEs-related anger, injustice, guilt, and insecurity) between TEs and PTSD symptoms. Community members in Hong Kong who had utilized a screening tool for PTSD and depressive symptoms were invited to complete a survey on exposure to stressful events and event-based rumination. A total of 10,110 individuals completed the survey. Hierarchical regression analysis showed that rumination, TEs, and SLEs were among the significant predictors for PTSD symptoms (all P < 0.001), accounting for 32% of the variance. For depression, rumination, SLEs, and PEs were among the significant predictors (all P < 0.001), explaining 24.9% of the variance. Two-way analysis of variance of different recent and prior TEs showed significant dose-effect relationships. The effect of recent TEs on PTSD symptoms was potentiated by prior TEs (P = 0.005). COVID-19 PEs and prior TEs additively contributed to PTSD symptoms, with no significant interaction (P = 0.94). Meanwhile, recent TEs were also potentiated by SLEs (P = 0.002). The effects of TEs on PTSD symptoms were mediated by rumination (β = 0.38, standard error = 0.01, 95% confidence interval: 0.36 to 0.41), with 40.4% of the total effect explained. All 4 rumination subtypes were significant mediators. Prior and ongoing TEs, PEs, and SLEs cumulatively exacerbated PTSD and depressive symptoms. The role of event-based rumination and their interventions should be prioritized for future research.

A Comprehensive Analysis of Weakly-Supervised Semantic Segmentation in Different Image Domains

Recently proposed methods for weakly-supervised semantic segmentation have achieved impressive performance in predicting pixel classes despite being trained with only image labels which lack positional information. Because image annotations are cheaper and quicker to generate, weak supervision is more practical than full supervision for training segmentation algorithms. These methods have been predominantly developed to solve the background separation and partial segmentation problems presented by natural scene images and it is unclear whether they can be simply transferred to other domains with different characteristics, such as histopathology and satellite images, and still perform well. This paper evaluates state-of-the-art weakly-supervised semantic segmentation methods on natural scene, histopathology, and satellite image datasets and analyzes how to determine which method is most suitable for a given dataset. Our experiments indicate that histopathology and satellite images present a different set of problems for weakly-supervised semantic segmentation than natural scene images, such as ambiguous boundaries and class co-occurrence. Methods perform well for datasets they were developed on, but tend to perform poorly on other datasets. We present some practical techniques for these methods on unseen datasets and argue that more work is needed for a generalizable approach to weakly-supervised semantic segmentation. Our full code implementation is available on GitHub: https://github.com/lyndonchan/wsss-analysis.

Abstract 3621: Characterization of BRAF mutation in Chinese colorectal cancer patients using next generation sequencing

Abstract Objective: BRAF alterations were characterized into three classes according to their dependency with upstream gene alterations in activating the MARK signaling pathway, giving clinical doctors clues of inhibitors selection for therapy. However, the characterization of BRAF mutation in Chinese colorectal cancer patients was unknown. Methods: Colorectum tissue samples were collected from 2155 patients and performed the next generation sequencing (NGS) using a validated commercial panel containing essential genes in RAS-RAF-MARK and PI3K-AKT-mTOR pathways. BRAF mutations were characterized into Class I (kinase-activated, codon 600), II (kinase-activated, non-codon 600) and III (kinase-impaired). Results: There were 9031 pathogenic and likely pathogenic somatic mutations were detected in 2137 out of 2155 samples, among which 149 (7.0%), 1048 (48.6%), 82 (3.8%), 2 (0.1%), 8 (0.4%), 362 (16.8%), 90 (4.2%) were found caring BRAF, KRAS, NRAS, HRAS, EGFR, PIK3CA and PTEN mutations, respectively. A total of 19 unique BRAF mutations were detected. The Class I (BRAFV600E), Class II and Class III accounted for 76.0%, 6.7% and 13.3% of all BRAF mutations. Higher incidence of co-occurrence of Class III mutations and RAS mutations were observed in our study, which might be due to their RAS-dependent way in activating downstream signaling pathway. Class III patients might be more sensitive to EGFR inhibitor and receive more benefits from survival than other classes. Conclusions: In our study, the mutation type, frequency, classification of BRAF mutations and co-occurrence with BRAF and other driver gene mutations were explored in a large cohort of Chinese colorectal cancer patients, which might help to suggest proper therapy methods. Citation Format: Hui Wang, Hao Wu, Yifan Zhou, Bei Zhang, Danpeng Sun, LIihua Jiang, Hui Chen, Yuezong Bai, Zhe Zhang. Characterization of BRAF mutation in Chinese colorectal cancer patients using next generation sequencing [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3621.

Class-Guided Feature Decoupling Network for Airborne Image Segmentation

Contextual information has been demonstrated to be helpful for airborne image segmentation. However, most of the previous works focus on the exploitation of spatially contextual information, which is difficult to segment isolated objects, mainly surrounded by uncorrelated objects. To alleviate this issue, we attempt to take advantage of the co-occurrence relations between different classes of objects in the scene. Especially, similar to other works, convolutional features are first extracted to capture the spatially contextual information. Then, a feature decoupling module is designed to encode the class co-occurrence relations into the convolutional features; thus, the most discriminative features can be decoupled. Finally, the segmentation result is inferred from the decoupled features. The whole process is integrated to form an end-to-end network, named class-guided feature decoupling network (CGFDN). Experimental results on two widely used benchmark data sets show that CGFDN obtains competitive results (>90% overall accuracy (OA) on 5-cm-resolution Potsdam and >91% OA on 9-cm-resolution Vaihingen) in comparison with several state-of-the-art models.

Determination of Antibiotic Resistance Profile and Virulence Genes in Escherichia coli Isolates from Palestinian Patients

Introduction: Escherichia coli (E. coli) is considered one of the most frequent intestinal and extraintestinal pathogen.&#x0D; Methods: A total of 49 isolates of E. coli were collected from different clinical samples, from different hospitals in Northern West Bank-Palestine, during January-March 2019.&#x0D; Aims: To detect the distribution of Type III secretion system (T3SS) genes, class 1, 2 and 3 integrons, virulence factors (fyuA, papGIII, iutA and sfa⁄foc) using multiplex PCR and antibiotic resistance using disc diffusion method.&#x0D; Results: In this study, E. coli isolates showed high resistance rate against different types of antibiotics and 71.4% of the isolates were multidrug resistant (MDR). Only class 1 integron was detected in these isolates with prevalence 57%, and 65.7% of MDR isolates carried integron genes. The prevalence of T3SS genes was 0.0%. In addition, results of this study showed that the prevalence of virulence genes papGIII, sfa⁄foc, fyuA and iutA was 4.1%, 40%, 64%, and 79.6%, respectively.&#x0D; Conclusions: The isolates of E. coli showed high resistance rate against different types of antibiotics. The co-occurrence of class 1 integrons and antimicrobial resistance genes in these isolates is an additional threat for spread of the antimicrobial resistance traits which may further complicate future strategies for treatment the infections caused by this pathogen. In addition, E. coli isolated from Palestinian patients showed one or more virulence factors that could increase their pathogenic potential.

Multi-label Learning with Missing Labels Using Mixed Dependency Graphs

This work focuses on the problem of multi-label learning with missing labels (MLML), which aims to label each test instance with multiple class labels given training instances that have an incomplete/partial set of these labels (i.e., some of their labels are missing). The key point to handle missing labels is propagating the label information from the provided labels to missing labels, through a dependency graph that each label of each instance is treated as a node. We build this graph by utilizing different types of label dependencies. Specifically, the instance-level similarity is served as undirected edges to connect the label nodes across different instances and the semantic label hierarchy is used as directed edges to connect different classes. This base graph is referred to as the mixed dependency graph, as it includes both undirected and directed edges. Furthermore, we present another two types of label dependencies to connect the label nodes across different classes. One is the class co-occurrence, which is also encoded as undirected edges. Combining with the above base graph, we obtain a new mixed graph, called mixed graph with co-occurrence (MG-CO). The other is the sparse and low rank decomposition of the whole label matrix, to embed high-order dependencies over all labels. Combining with the base graph, the new mixed graph is called as MG-SL (mixed graph with sparse and low rank decomposition). Based on MG-CO and MG-SL, we further propose two convex transductive formulations of the MLML problem, denoted as MLMG-CO and MLMG-SL respectively. In both formulations, the instance-level similarity is embedded through a quadratic smoothness term, while the semantic label hierarchy is used as a linear constraint. In MLMG-CO, the class co-occurrence is also formulated as a quadratic smoothness term, while the sparse and low rank decomposition is incorporated into MLMG-SL, through two additional matrices (one is assumed as sparse, and the other is assumed as low rank) and an equivalence constraint between the summation of this two matrices and the original label matrix. Interestingly, two important applications, including image annotation and tag based image retrieval, can be jointly handled using our proposed methods. Experimental results on several benchmark datasets show that our methods lead to significant improvements in performance and robustness to missing labels over the state-of-the-art methods.

Matching biomedical ontologies based on formal concept analysis

BackgroundThe goal of ontology matching is to identify correspondences between entities from different yet overlapping ontologies so as to facilitate semantic integration, reuse and interoperability. As a well developed mathematical model for analyzing individuals and structuring concepts, Formal Concept Analysis (FCA) has been applied to ontology matching (OM) tasks since the beginning of OM research, whereas ontological knowledge exploited in FCA-based methods is limited. This motivates the study in this paper, i.e., to empower FCA with as much as ontological knowledge as possible for identifying mappings across ontologies.MethodsWe propose a method based on Formal Concept Analysis to identify and validate mappings across ontologies, including one-to-one mappings, complex mappings and correspondences between object properties. Our method, called FCA-Map, incrementally generates a total of five types of formal contexts and extracts mappings from the lattices derived. First, the token-based formal context describes how class names, labels and synonyms share lexical tokens, leading to lexical mappings (anchors) across ontologies. Second, the relation-based formal context describes how classes are in taxonomic, partonomic and disjoint relationships with the anchors, leading to positive and negative structural evidence for validating the lexical matching. Third, the positive relation-based context can be used to discover structural mappings. Afterwards, the property-based formal context describes how object properties are used in axioms to connect anchor classes across ontologies, leading to property mappings. Last, the restriction-based formal context describes co-occurrence of classes across ontologies in anonymous ancestors of anchors, from which extended structural mappings and complex mappings can be identified.ResultsEvaluation on the Anatomy, the Large Biomedical Ontologies, and the Disease and Phenotype track of the 2016 Ontology Alignment Evaluation Initiative campaign demonstrates the effectiveness of FCA-Map and its competitiveness with the top-ranked systems. FCA-Map can achieve a better balance between precision and recall for large-scale domain ontologies through constructing multiple FCA structures, whereas it performs unsatisfactorily for smaller-sized ontologies with less lexical and semantic expressions.ConclusionsCompared with other FCA-based OM systems, the study in this paper is more comprehensive as an attempt to push the envelope of the Formal Concept Analysis formalism in ontology matching tasks. Five types of formal contexts are constructed incrementally, and their derived concept lattices are used to cluster the commonalities among classes at lexical and structural level, respectively. Experiments on large, real-world domain ontologies show promising results and reveal the power of FCA.

A Markov random field integrating spectral dissimilarity and class co-occurrence dependency for remote sensing image classification optimization

This paper develops a novel Markov Random Field (MRF) model for edge-preserving spatial regularization of classification maps. MRF methods based on the uniform smoothness lead to oversmoothed solutions. In contrast, MRF methods which take care of local spectral or gradient discontinuities, lead to unexpected object particles around boundaries. To solve these key problems, our developed MRF method first establishes a spatial energy function integrating local spectral dissimilarity to smooth the initial classification map while preserving object boundaries. Second, a new anisotropic spatial energy function integrating the class co-occurrence dependency is constructed to regularize pixels around object boundaries. The effectiveness of the method is tested using a series of remote sensing data sets. The obtained results indicate that the method can avoid oversmoothing and significantly improve the classification accuracy with regards to traditional MRF classification models and some other state-of-the-art methods.

On using landscape metrics for landscape similarity search

Landscape similarity search involves finding landscapes from among a large collection that are similar to a query landscape. An example of such collection is a large land cover map subdivided into a grid of smaller local landscapes, a query is a local landscape of interest, and the task is to find other local landscapes within a map which are perceptually similar to the query. Landscape search and the related task of pattern-based regionalization, requires a measure of similarity – a function which quantifies the level of likeness between two landscapes. The standard approach is to use the Euclidean distance between vectors of landscape metrics derived from the two landscapes, but no in-depth analysis of this approach has been conducted. In this paper we investigate the performance of different implementations of the standard similarity measure. Five different implementations are tested against each other and against a control similarity measure based on histograms of class co-occurrence features and the Jensen–Shannon divergence. Testing consists of a series of numerical experiments combined with visual assessments on a set of 400 3km-scale landscapes. Based on the cases where visual assessment provides definitive answer, we have determined that the standard similarity measure is sensitive to the way landscape metrics are normalized and, additionally, to whether weights aimed at controlling the relative contribution of landscape composition vs. configuration are used. The standard measure achieves the best performance when metrics are normalized using their extreme values extracted from all possible landscapes, not just the landscapes in the given collection, and when weights are assigned so the combined influence of composition metrics on the similarity value equals the combined influence of configuration metrics. We have also determined that the control similarity measure outperforms all implementations of the standard measure.

Detecting and quantifying extended landscape structure with spatial co-occurrence surfaces

The attribute adjacency matrix is a fundamental component of many metrics used to characterize landscape heterogeneity from land cover land use maps. Since it quantifies adjacent pixel class co-occurrence, it is unsuited to detect broader-scale structure in land cover maps. This paper proposes a generalization of the adjacency matrix concept by incorporating lag distances into class co-occurrence estimation. The spectrum of spatial structure is presented in the form of a spatial co-occurrence surface. These surfaces can provide a wealth of information on landscape structure including the size and spatial distribution of patches of a single class as well as inter-class spatial associations.

Mutation Profiling of Therapy-Related Myeloid Neoplasms Using Next-Generation Sequencing Demonstrates Distinct Profiles from De Novo Disease

Therapy-related Myeloid Neoplasms (tMN) comprise a poor risk subset of myelodysplastic syndromes and acute myelogenous leukemia, are increasing in incidence, and represent a serious complication following treatment for primary malignancies. In our previous study of 11 genes in 38 tMN patient samples, the data suggested that the mutational spectrum of tMN was distinct from de novo myeloid malignancies. To confirm this finding and to refine the tMN mutation profile, we investigated the mutation profile in samples from 88 patients and 28 genes using Sanger and next-generation sequencing approaches. We performed amplification using RainDance microfluidic PCR, followed by HiSeq next-generation sequencing. Mutations were identified using a modified pipeline for SNP calling employing variant detection software programs.Our study cohort included 88 patients, 71 of whom had complete clinical data for analyses. Patients had a history of epithelial and hematologic malignancies (³2 malignancies n=11; breast n=9; colorectal n=5; head and neck n=4; genital-urinary n=6; lung n=1; lymphoma n=25; melanoma n=2; ovarian n=1; sarcoma n=2; other, n=5). Treatment of primary cancers included chemotherapy alone (n=27), radiation alone (n=8), autologous stem cell transplant (n=11), or chemotherapy plus radiation (n=25). The median latency time between primary malignancy treatment and tMN diagnosis was 5.7yrs (range, 0.7 - 30.9 yrs). Median age at tMN diagnosis was 64yrs (range, 26 - 85 yrs). International Prognostic Scoring System (IPSS) risk group for MDS at tMN diagnosis were Low risk (n=8), Int-1 (n=11), Int-2 (n=30), High risk (n=9).We identified somatic mutations in 56 of 88 (64%) patients (83 patients were evaluated by next-generation sequencing and 5 by Sanger sequencing only). Mutations in TP53 were most common and were detected in 27/88 patients (30.7%), followed by mutations in TET2 in 12/88 (13.6%), DNMT3A in 9/88 (10.2%), NRAS in 8/83 (9.6%), KRAS in 5/83 (6.0%), and KIT in 5/83 (6.0%). Gene mutations detected at lower frequencies included those in ASXL1 in 5/88 (5.7%), RUNX1 in 2/83 (2.4%), EZH2 in 1/88 (1.1%), and SF3B1 in 1/88 (1.1%). Of the 58 patients with complete sequencing and FISH data, 4 patients exhibited biallelic somatic TP53 mutations and 3 patients had TP53 mutation combined with del 17p TP53 loss, demonstrating that 7 of 58 evaluable patients (12.1%) experienced biallelic loss of TP53. We also identified biallelic mutations in TET2 and DNMT3A in 2 separate patients. 25 patients had 2 or more concurrent somatic mutations. The highest number of co-occurring mutations in one patient was 5 mutations; 12 patients had 2 somatic mutations. The most common co-occurrence was TP53 and TET2, which was observed in 5 patients. All 5 ASXL1 mutations co-occurred with additional mutations.By analyzing variant allele frequencies (VAFs) in patients with multiple mutations, we observed that some tMN patients harbored multiple clones with distinct VAFs. This observation was also supported by the co-occurrence of typical class I driver mutations in the same patient, (e.g. KRAS 6% and NRAS 21% VAF; NRAS 9% and KIT 34%; NRAS 26% and KIT 9% in individual patients). The allele frequency data also suggested that ASXL1 is likely an early occurring mutation as the VAF was higher than for other co-occurring mutations (mean VAF ASXL1 50%, other co-occurring genes 23.5%, p<.05 t-test).Because of previous reports on the prognostic significance of point mutations in myeloid malignancies (e.g. TP53 in MDS and TET2 in AML), we tested the impact of individual mutations on prognosis. TP53 mutation or loss was associated with worse prognosis in tMN (OS 17.6 vs 25.2 mos, n=72, p<.11 log-rank test) (Fig A). TET2 mutation and KRAS or NRAS mutations did not predict for a difference in prognosis, although analysis was limited by cohort size. TP53 mutation was also associated with del 5q / monosomy 5 (p<.0001, Chi-square test, n=51).Our data reveal that tMNs display distinct mutation profile compared to de novo disease (Fig B). TP53 mutations and loss are the most common abnormalities and predict for adverse outcome. Epigenetic modifier mutations also occur in tMNs and can serve as disease-initiating mutations. Collectively our results demonstrate that characterizing these mutation profiles can enhance our understanding of disease mechanisms in tMNs and may guide the development of future therapies for these difficult to treat disorders. [Display omitted] DisclosuresSekeres:Celgene Corp.: Membership on an entity’s Board of Directors or advisory committees; Amgen: Membership on an entity’s Board of Directors or advisory committees; Boehringer Ingelheim: Membership on an entity’s Board of Directors or advisory committees.

Smoothing Parameter Estimation for Markov Random Field Classification of non-Gaussian Distribution Image

Abstract. In the context of remote sensing image classification, Markov random fields (MRFs) have been used to combine both spectral and contextual information. The MRFs use a smoothing parameter to balance the contribution of the spectral versus spatial energies, which is often defined empirically. This paper proposes a framework to estimate the smoothing parameter using the probability estimates from support vector machines and the spatial class co-occurrence distribution. Furthermore, we construct a spatially weighted parameter to preserve the edges by using seven different edge detectors. The performance of the proposed methods is evaluated on two hyperspectral datasets recorded by the AVIRIS and ROSIS and a simulated ALOS PALSAR image. The experimental results demonstrated that the estimated smoothing parameter is optimal and produces a classified map with high accuracy. Moreover, we found that the Canny-based edge probability map preserved the contours better than others.