Non-overlapping Classes Research Articles

Single-Cell Analysis of Rohon-Beard Neurons Implicates Fgf Signaling in Axon Maintenance and Cell Survival.

Peripheral sensory neurons are a critical part of the nervous system that transmit a multitude of sensory stimuli to the central nervous system. During larval and juvenile stages in zebrafish, this function is mediated by Rohon-Beard somatosensory neurons (RBs). RBs are optically accessible and amenable to experimental manipulation, making them a powerful system for mechanistic investigation of sensory neurons. Previous studies provided evidence that RBs fall into multiple subclasses; however, the number and molecular makeup of these potential RB subtypes have not been well defined. Using a single-cell RNA sequencing (scRNA-seq) approach, we demonstrate that larval RBs in zebrafish fall into three, largely nonoverlapping classes of neurons. We also show that RBs are molecularly distinct from trigeminal neurons in zebrafish. Cross-species transcriptional analysis indicates that one RB subclass is similar to a mammalian group of A-fiber sensory neurons. Another RB subclass is predicted to sense multiple modalities, including mechanical stimulation and chemical irritants. We leveraged our scRNA-seq data to determine that the fibroblast growth factor (Fgf) pathway is active in RBs. Pharmacological and genetic inhibition of this pathway led to defects in axon maintenance and RB cell death. Moreover, this can be phenocopied by treatment with dovitinib, an FDA-approved Fgf inhibitor with a common side effect of peripheral neuropathy. Importantly, dovitinib-mediated axon loss can be suppressed by loss of Sarm1, a positive regulator of neuronal cell death and axonal injury. This offers a molecular target for future clinical intervention to fight neurotoxic effects of this drug.

DeepAstroUDA: semi-supervised universal domain adaptation for cross-survey galaxy morphology classification and anomaly detection

Artificial intelligence methods show great promise in increasing the quality and speed of work with large astronomical datasets, but the high complexity of these methods leads to the extraction of dataset-specific, non-robust features. Therefore, such methods do not generalize well across multiple datasets. We present a universal domain adaptation method, DeepAstroUDA, as an approach to overcome this challenge. This algorithm performs semi-supervised domain adaptation (DA) and can be applied to datasets with different data distributions and class overlaps. Non-overlapping classes can be present in any of the two datasets (the labeled source domain, or the unlabeled target domain), and the method can even be used in the presence of unknown classes. We apply our method to three examples of galaxy morphology classification tasks of different complexities (three-class and ten-class problems), with anomaly detection: (1) datasets created after different numbers of observing years from a single survey (Legacy Survey of Space and Time mock data of one and ten years of observations); (2) data from different surveys (Sloan Digital Sky Survey (SDSS) and DECaLS); and (3) data from observing fields with different depths within one survey (wide field and Stripe 82 deep field of SDSS). For the first time, we demonstrate the successful use of DA between very discrepant observational datasets. DeepAstroUDA is capable of bridging the gap between two astronomical surveys, increasing classification accuracy in both domains (up to on the unlabeled data), and making model performance consistent across datasets. Furthermore, our method also performs well as an anomaly detection algorithm and successfully clusters unknown class samples even in the unlabeled target dataset.

The ‘Service Transformation Outline’ toolkit to assist in the design, comparison and replication of integrated care programmes

PurposeThe author is introducing a toolkit that can assist stakeholders to design, compare and replicate integrated care programmes, by making explicit their view on the transformations of care and cure services, according to a structured template. The purpose of this paper is to address this issue.Design/methodology/approachThe toolkit is made of two elements: a classification and a template. The author adopted a step-wise approach of semantic modelling to work out three layers of a classification in the domain of integration needs related to care and cure services. The third layer consists of 23 non-overlapping classes that fully cover that semantic domain. The classes are used to build a template to elicit the stakeholder's standpoint about the transformations involved in the deployment of a programme. The result is the «Outline» of the programme.FindingsSo far, in eight years the author applied the toolkit to 100+ programmes either to design, simulate or evaluate them, either to compare them to similar ones in the same or in different jurisdictions, and we refined the description of the classes according to that experience.Research limitations/implicationsThe objective of the toolkit is not to provide solutions, but to stimulate reflections on the transformations involved in a programme and their practical consequences in a precise context. In fact, the Outlines cannot be generalized: they are conceived to reflect the perspective of the stakeholders and thus are intrinsically subjective; in addition, they must be contingent, as they must depend on the local context in the particular timeframe. In case of similar initiatives in other localities, the Outlines must be filled in again by the local stakeholders; however, a subsequent comparison could help to explore similarities and motivated differences.Practical implicationsIdeally the stakeholders should use the Outline of the actual transformations in a programme as a reference to mediate between the principles and the methodologies provided by the «Models» and «Conceptual Frameworks» in the literature (e.g. Chronic Care Model, Rainbow Model, Development Model of Integrated Care Patient-Centered Medical Home and Maturity Models) and the deeper studies using the specific tools developed in their disciplines (e.g. on information modelling, process modelling, cost–benefit analysis and health technology assessment). The toolkit could have its role also in a multi-annual roadmap made of a sequence of programmes to cope with the urgent challenges on ageing, social changes and technological evolution, in synergy with regulations, budget, context maturity, critical success factors and local priorities.Originality/valueThe author argues that the approach of the structured Outline is unique, as the scoring mechanism to assess the relevance of the transformation within each class on the overall change brought by the whole programme into the health systems.

Learning in deep neural networks and brains with similarity-weighted interleaved learning

Understanding how the brain learns throughout a lifetime remains a long-standing challenge. In artificial neural networks (ANNs), incorporating novel information too rapidly results in catastrophic interference, i.e., abrupt loss of previously acquired knowledge. Complementary Learning Systems Theory (CLST) suggests that new memories can be gradually integrated into the neocortex by interleaving new memories with existing knowledge. This approach, however, has been assumed to require interleaving all existing knowledge every time something new is learned, which is implausible because it is time-consuming and requires a large amount of data. We show that deep, nonlinear ANNs can learn new information by interleaving only a subset of old items that share substantial representational similarity with the new information. By using such similarity-weighted interleaved learning (SWIL), ANNs can learn new information rapidly with a similar accuracy level and minimal interference, while using a much smaller number of old items presented per epoch (fast and data-efficient). SWIL is shown to work with various standard classification datasets (Fashion-MNIST, CIFAR10, and CIFAR100), deep neural network architectures, and in sequential learning frameworks. We show that data efficiency and speedup in learning new items are increased roughly proportionally to the number of nonoverlapping classes stored in the network, which implies an enormous possible speedup in human brains, which encode a high number of separate categories. Finally, we propose a theoretical model of how SWIL might be implemented in the brain.

Types of technical state in transport systems

The concepts of the type of technical state and a set of types of technical state are included in the basic set of concepts for technical diagnostics. Until now, there is a problem of assessing the necessary or sufficient capacity of a set of types of technical state. The technology for identifying the types of technical state, which is built on the sequential division of a set of defects into two non-overlapping classes, has significant pragmatic drawbacks. In this work, based on the systematization and formalization of the concepts of state, technical state and types of technical state, a pragmatic approach to the formation of a set of types of technical state is substantiated and analyzed. Within the framework of this approach, it is proposed to distinguish two time scales in diagnostic tasks – functional and degradation, to analyze objective and subjective factors in the technology of identifying types of technical states, and also to take into account the pragmatic aspect in the formation of a set of types of technical states. The formation of a fairly representative set of types of technical state will allow organizing optimal actions of the operating and maintenance personnel, adequate to the current technical state of the transport system, and minimizing the cost of restoring this state to normal.

Lorenz and Polarization Orderings of the Double-Pareto Lognormal Distribution and Other Size Distributions

Polarization indices such as the Foster-Wolfson index have been developed to measure the extent of clustering in a few classes with wide gaps between them in terms of income distribution. However, Zhang and Kanbur (2001) failed to empirically find clear differences between polarization and inequality indices in the measurement of intertemporal distributional changes. This paper addresses this ‘distinction' problem on the level of the respective underlying stochastic orders, the polarization order (PO) in distributions divided into two nonoverlapping classes and the Lorenz order (LO) of inequality in distributions. More specifically, this paper investigates whether a distribution F can be either more or less polarized than a distribution H in terms of the PO if F is more unequal than H in terms of the LO. Furthermore, this paper derives conditions for the LO and PO of the double-Pareto lognormal (dPLN) distribution. The derived conditions are applicable to sensitivity analyses of inequality and polarization indices with respect to distributional changes. From this application, a suggestion for appropriate two-class polarization indices is made.

An Iterative Approach to Stratification: Poverty at Regional Level in Italy

We develop an iterative procedure to generate stratification in non-overlapping classes of a population based on a one-dimensional variable, namely, the Italian income. The procedure works under the assumption that the income distribution is generated by a log-normal mixture of homogeneous income groups. The number of income groups is not determined a priori but rather endogenously by the iterative procedure that stops when the solution to a specific second-degree polynomial equation does not exist. We apply the approach to study the heterogeneity of Italian incomes and the poor population at regional level in the years 2005, 2010, and 2015. The cross-regional comparisons show differences in inequality and stratification dynamics while comparisons over time show the evolution of the classes.

Neoantigen Dissimilarity to the Self-Proteome Predicts Immunogenicity and Response to Immune Checkpoint Blockade.

Despite improved methods for MHC affinity prediction, the vast majority of computationally predicted tumor neoantigens are not immunogenic experimentally, indicating that high-quality neoantigens are beyond current algorithms to discern. To enrich for neoantigens with the greatest likelihood of immunogenicity, we developed an analytic method to parse neoantigen quality through rational biological criteria across five clinical datasets for 318 cancer patients. We explored four quality metrics, including analysis of dissimilarity to the non-mutated proteome that was predictive of peptide immunogenicity. In patient tumors, neoantigens with high dissimilarity were unique, enriched for hydrophobic sequences, and correlated with survival after PD-1 checkpoint therapy in patients with non-small cell lung cancer independent of predicted MHC affinity. We incorporated our neoantigen quality analysis methodology into an open-source tool, antigen.garnish, to predict immunogenic peptides from bulk computationally predicted neoantigens for which the immunogenic "hit rate" is currently low.

An Intelligent Computer-Aided Scheme for Classifying Multiple Skin Lesions

Skin diseases cases are increasing on a daily basis and are difficult to handle due to the global imbalance between skin disease patients and dermatologists. Skin diseases are among the top 5 leading cause of the worldwide disease burden. To reduce this burden, computer-aided diagnosis systems (CAD) are highly demanded. Single disease classification is the major shortcoming in the existing work. Due to the similar characteristics of skin diseases, classification of multiple skin lesions is very challenging. This research work is an extension of our existing work where a novel classification scheme is proposed for multi-class classification. The proposed classification framework can classify an input skin image into one of the six non-overlapping classes i.e., healthy, acne, eczema, psoriasis, benign and malignant melanoma. The proposed classification framework constitutes four steps, i.e., pre-processing, segmentation, feature extraction and classification. Different image processing and machine learning techniques are used to accomplish each step. 10-fold cross-validation is utilized, and experiments are performed on 1800 images. An accuracy of 94.74% was achieved using Quadratic Support Vector Machine. The proposed classification scheme can help patients in the early classification of skin lesions.

Case-control analysis identifies shared properties of rare germline variation in cancer predisposing genes.

Along with traditional effects of aging and carcinogen exposure-inherited DNA variation has substantial contribution to cancer risk. Extraordinary progress made in analysis of common variation with GWAS methodology does not provide sufficient resolution to understand rare variation. To fulfill missing classification for rare germline variation we assembled dataset of whole exome sequences from>2000 patients (selected cases tested negative for candidate genes and unselected cases) with different types of cancers (breast cancer, colon cancer, and cutaneous and ocular melanomas) matched to more than 7000 non-cancer controls and analyzed germline variation in known cancer predisposing genes to identify common properties of disease-associated DNA variation and aid the future searches for new cancer susceptibility genes. Cancer predisposing genes were divided into non-overlapping classes according to the mode of inheritance of the related cancer syndrome or known tumor suppressor activity. Out of all classes only genes linked to dominant syndromes presented significant rare germline variants enrichment in cases. Separate analysis of protein-truncating and missense variation in this list of genes confirmed significant prevalence of protein-truncating variants in cases only in loss-of-function tolerant genes (pLI < 0.1), while ultra-rare missense variants were significantly overrepresented in cases only in constrained genes (pLI > 0.9). In addition to findings in genetically enriched cases, we observed significant burden of rare variation in unselected cases, suggesting substantial role of inherited variation even in relatively late cancer manifestation. Taken together, our findings provide reference for distribution and types of DNA variation underlying inherited predisposition to some common cancer types.

3D crystal structure identification using fuzzy neural networks

The problem of recognizing nano-scale images of lattice projections comes down to identification of crystal lattice structure. The paper considers two types of fuzzy neural networks that can be used for tackling the problem at hand: the Takagi-Sugeno-Kang model and Mamdani-Zadeh model (the latter being a modification of the Wang-Mendel fuzzy neural network). We offer a threestage neural network learning process. In the first two stages crystal lattices are grouped in non-overlapping classes, and lattices belonging to overlapping classes are recognized at the third stage. In the research, we thoroughly investigate the applicability of the neural net models to structure identification of 3D crystal lattices.

Distribution of type I corticotropin-releasing factor (CRF1) receptors on GABAergic neurons within the basolateral amygdala.

The neuropeptide corticotropin-releasing factor (CRF) plays a critical role in mediating anxiety-like responses to stressors, and dysfunction of the CRF system has been linked to the etiology of several psychiatric disorders. Extra-hypothalamic CRF can also modulate learning and memory formation, including amygdala-dependent learning. The basolateral nucleus of the amygdala (BLA) contains dense concentrations of CRF receptors, yet the distribution of these receptors on specific neuronal subtypes within the BLA has not been characterized. Here, we quantified the expression of CRF receptors on three nonoverlapping classes of GABAergic interneurons: those containing the calcium-binding protein parvalbumin (PV), and those expressing the neuropeptides somatostatin (SOM) or cholecystokinin (CCK). While the majority of PV+ neurons and roughly half of CCK+ neurons expressed CRF receptors, they were expressed to a much lesser extent on SOM+ interneurons. Knowledge of the distribution of CRF receptors within the BLA can provide insight into how manipulations of the CRF system modulate fear and anxiety-like behaviors.

Analysis of feature selection measures for text categorisation

The curse of dimensionality has made dimension reduction an essential step in text categorisation. Feature selection is an approach for dimension reduction. In this paper an analysis on feature selection measures for text categorisation is performed. Under the unsupervised approach document frequency and under the supervised approach chi-square, odds ratio, mutual information, and information gain are considered for analysis. They are considered here because they are the widely used and effective measures. Analysis of these measures is performed using the 20 newsgroups dataset. Twenty newsgroups dataset consists of closely related categories as well as highly unrelated categories. Certain categories of 20 newsgroups dataset are selected and organised into three groups of overlapping (highly related) classes, non-overlapping (highly unrelated) classes and combination of overlapping and non-overlapping classes. Feature selection and subsequent classification is applied to the three groups separately and the classification performance is studied based on the feature selection measures. The noticeable behaviour was with odds ratio measure in that it performed well for non-overlapping group and overlapping groups considered separately and was poorer in performance for the group containing both overlapping and non-overlapping categories. Remaining measures showed consistent behaviour for all the three groups. Classification was achieved using support vector machine classifier. The performance comparisons of different measures on different groups are presented in terms of micro-F1 and macro-F1.

Analysis of feature selection measures for text categorisation

The curse of dimensionality has made dimension reduction an essential step in text categorisation. Feature selection is an approach for dimension reduction. In this paper an analysis on feature selection measures for text categorisation is performed. Under the unsupervised approach document frequency and under the supervised approach chi-square, odds ratio, mutual information, and information gain are considered for analysis. They are considered here because they are the widely used and effective measures. Analysis of these measures is performed using the 20 newsgroups dataset. Twenty newsgroups dataset consists of closely related categories as well as highly unrelated categories. Certain categories of 20 newsgroups dataset are selected and organised into three groups of overlapping (highly related) classes, non-overlapping (highly unrelated) classes and combination of overlapping and non-overlapping classes. Feature selection and subsequent classification is applied to the three groups separately and the classification performance is studied based on the feature selection measures. The noticeable behaviour was with odds ratio measure in that it performed well for non-overlapping group and overlapping groups considered separately and was poorer in performance for the group containing both overlapping and non-overlapping categories. Remaining measures showed consistent behaviour for all the three groups. Classification was achieved using support vector machine classifier. The performance comparisons of different measures on different groups are presented in terms of micro-F1 and macro-F1.

DNA cytosine hydroxymethylation levels are distinct among non-overlapping classes of peripheral blood leukocytes

BackgroundPeripheral blood leukocytes are the most commonly used surrogates to study epigenome-induced risk and epigenomic response to disease-related stress. We considered the hypothesis that the various classes of peripheral leukocytes differentially regulate the synthesis of 5-methylcytosine (5mCG) and its removal via Ten-Eleven Translocation (TET) dioxygenase catalyzed hydroxymethylation to 5-hydroxymethylcytosine (5hmCG), reflecting their responsiveness to environment. Although it is known that reductions in TET1 and/or TET2 activity lead to the over-proliferation of various leukocyte precursors in bone marrow and in development of chronic myelomonocytic leukemia and myeloproliferative neoplasms, the role of 5mCG hydroxymethylation in peripheral blood is less well studied. ResultsWe developed simplified protocols to rapidly and reiteratively isolate non-overlapping leukocyte populations from a single small sample of fresh or frozen whole blood. Among peripheral leukocyte types we found extreme variation in the levels of transcripts encoding proteins involved in cytosine methylation (DNMT1, 3A, 3B), the turnover of 5mC by demethylation (TET1, 2, 3), and DNA repair (GADD45A, B, G) and in the global and gene-region-specific levels of DNA 5hmCG (CD4+ T cells≫CD14+ monocytes>CD16+ neutrophils>CD19+ B cells>CD56+ NK cells>Siglec8+ eosinophils>CD8+ T cells). ConclusionsOur data taken together suggest a potential hierarchy of responsiveness among classes of leukocytes with CD4+, CD8+ T cells and CD14+ monocytes being the most distinctly poised for a rapid methylome response to physiological stress and disease.

Albayzín-2014 evaluation: audio segmentation and classification in broadcast news domains

Audio segmentation is important as a pre-processing task to improve the performance of many speech technology tasks and, therefore, it has an undoubted research interest. This paper describes the database, the metric, the systems and the results for the Albayzin-2014 audio segmentation campaign. In contrast to previous evaluations where the task was the segmentation of non-overlapping classes, Albayzin-2014 evaluation proposes the delimitation of the presence of speech, music and/or noise that can be found simultaneously. The database used in the evaluation was created by fusing different media and noises in order to increase the difficulty of the task. Seven segmentation systems from four different research groups were evaluated and combined. Their experimental results were analyzed and compared with the aim of providing a benchmark and showing up the promising directions in this field.

Transcriptional profiling at whole population and single cell levels reveals somatosensory neuron molecular diversity.

The somatosensory nervous system is critical for the organism's ability to respond to mechanical, thermal, and nociceptive stimuli. Somatosensory neurons are functionally and anatomically diverse but their molecular profiles are not well-defined. Here, we used transcriptional profiling to analyze the detailed molecular signatures of dorsal root ganglion (DRG) sensory neurons. We used two mouse reporter lines and surface IB4 labeling to purify three major non-overlapping classes of neurons: 1) IB4(+)SNS-Cre/TdTomato(+), 2) IB4(-)SNS-Cre/TdTomato(+), and 3) Parv-Cre/TdTomato(+) cells, encompassing the majority of nociceptive, pruriceptive, and proprioceptive neurons. These neurons displayed distinct expression patterns of ion channels, transcription factors, and GPCRs. Highly parallel qRT-PCR analysis of 334 single neurons selected by membership of the three populations demonstrated further diversity, with unbiased clustering analysis identifying six distinct subgroups. These data significantly increase our knowledge of the molecular identities of known DRG populations and uncover potentially novel subsets, revealing the complexity and diversity of those neurons underlying somatosensation.

Scorpion image segmentation system

Death as a result of scorpion sting has been a major public health problem in developing countries. Despite the high rate of death as a result of scorpion sting, little report exists in literature of intelligent device and system for automatic detection of scorpion. This paper proposed a digital image processing approach based on the floresencing characteristics of Scorpion under Ultra-violet (UV) light for automatic detection and identification of scorpion. The acquired UV-based images undergo pre-processing to equalize uneven illumination and colour space channel separation. The extracted channels are then segmented into two non-overlapping classes. It has been observed that simple thresholding of the green channel of the acquired RGB UV-based image is sufficient for segmenting Scorpion from other background components in the acquired image. Two approaches to image segmentation have also been proposed in this work, namely, the simple average segmentation technique and K-means image segmentation. The proposed algorithm has been tested on over 40 UV scorpion images obtained from different part of the world and results obtained show an average accuracy of 97.7% in correctly classifying the pixel into two non-overlapping clusters. The proposed 1system will eliminate the problem associated with some of the existing manual approaches presently in use for scorpion detection.

Functional differences between neurochemically defined populations of inhibitory interneurons in the rat spinal dorsal horn

In order to understand how nociceptive information is processed in the spinal dorsal horn we need to unravel the complex synaptic circuits involving interneurons, which constitute the vast majority of the neurons in laminae I–III. The main limitation has been the difficulty in defining functional populations among these cells. We have recently identified 4 non-overlapping classes of inhibitory interneuron, defined by expression of galanin, neuropeptide Y (NPY), neuronal nitric oxide synthase (nNOS) and parvalbumin, in the rat spinal cord. In this study we demonstrate that these form distinct functional populations that differ in terms of sst2A receptor expression and in their responses to painful stimulation. The sst2A receptor was expressed by nearly all of the nNOS- and galanin-containing inhibitory interneurons but by few of those with NPY and none of the parvalbumin cells. Many galanin- and NPY-containing cells exhibited phosphorylated extracellular signal-regulated kinases (pERK) after mechanical, thermal or chemical noxious stimuli, but very few nNOS-containing cells expressed pERK after any of these stimuli. However, many nNOS-positive inhibitory interneurons up-regulated Fos after noxious thermal stimulation or injection of formalin, but not after capsaicin injection. Parvalbumin cells did not express either activity-dependent marker following any of these stimuli. These results suggest that interneurons belonging to the NPY, nNOS and galanin populations are involved in attenuating pain, and for NPY and nNOS cells this is likely to result from direct inhibition of nociceptive projection neurons. They also suggest that the nociceptive inputs to the nNOS cells differ from those to the galanin and NPY populations.

Neurochemical Characterisation of Lamina II Inhibitory Interneurons that Express GFP in the PrP-GFP Mouse

BackgroundInhibitory interneurons in the superficial dorsal horn play important roles in modulating sensory transmission, and these roles are thought to be performed by distinct functional populations. We have identified 4 non-overlapping classes among the inhibitory interneurons in the rat, defined by the presence of galanin, neuropeptide Y, neuronal nitric oxide synthase (nNOS) and parvalbumin. The somatostatin receptor sst2A is expressed by ~50% of the inhibitory interneurons in this region, and is particularly associated with nNOS- and galanin-expressing cells. The main aim of the present study was to test whether a genetically-defined population of inhibitory interneurons, those expressing green fluorescent protein (GFP) in the PrP-GFP mouse, belonged to one or more of the neurochemical classes identified in the rat.ResultsThe expression of sst2A and its relation to other neurochemical markers in the mouse was similar to that in the rat, except that a significant number of cells co-expressed nNOS and galanin. The PrP-GFP cells were entirely contained within the set of inhibitory interneurons that possessed sst2A receptors, and virtually all expressed nNOS and/or galanin. GFP was present in ~3-4% of neurons in the superficial dorsal horn, corresponding to ~16% of the inhibitory interneurons in this region. Consistent with their sst2A-immunoreactivity, all of the GFP cells were hyperpolarised by somatostatin, and this was prevented by administration of a selective sst2 receptor antagonist or a blocker of G-protein-coupled inwardly rectifying K+ channels.ConclusionsThese findings support the view that neurochemistry provides a valuable way of classifying inhibitory interneurons in the superficial laminae. Together with previous evidence that the PrP-GFP cells form a relatively homogeneous population in terms of their physiological properties, they suggest that these neurons have specific roles in processing sensory information in the dorsal horn.