Context-specific Features Research Articles

Applications of machine learning & Internet of Things for outdoor air pollution monitoring and prediction: A systematic literature review

According to the World Health Organization (WHO), air pollution kills seven million people every year. Outdoor air pollution is a major environmental health problem affecting low, middle, and high-income countries. In the past few years, the research community has explored IoT-enabled machine learning applications for outdoor air pollution prediction. The general objective of this paper is to systematically review applications of machine learning and Internet of Things (IoT) for outdoor air pollution prediction and the combination of monitoring sensors and input features used. Two research questions were formulated for this review. 1086 publications were collected in the initial PRISMA stage. After the screening and eligibility phases, 37 papers were selected for inclusion. A cost-based analysis was conducted on the findings to highlight high-cost monitoring, low-cost IoT and hybrid enabled prediction. Three methods of prediction were identified: time series, feature-based and spatio-temporal. This review’s findings identify major limitations in applications found in the literature, namely lack of coverage, lack of diversity of data and lack of inclusion of context-specific features. This review proposes directions for future research and underlines practical implications in healthcare, urban planning, global synergy and smart cities.

Requirements for a dashboard optimized for melanoma patient care through user-centered context exploration

For time-sensitive treatment of a patient with malignant melanoma, physicians must obtain a rapid overview of the patient’s status. This study aimed to analyze context-specific features and processes at the point of care to derive requirements for a dashboard granting more straightforward access to information. The Think-Aloud method, contextual inquiries, and interviews were performed with physicians from the Department of Dermatology at the University Hospital Essen in Germany. The user statements and observations that were obtained were grouped and categorized using an affinity diagram. Based on the derived subjects, requirements were defined, confirmed, and prioritized. The resulting affinity diagram revealed four topics of importance at the point of care. These topics are “Identifying and Processing the Important”, a comprehensive “Patient Record”, tasks and challenges in the “Clinical Routine”, and interactions and experiences with the available “Systems”. All aspects have been reflected in 135 requirements for developing context- and indication-specific patient dashboards. Our work has elucidated the most important aspects to consider when designing a dashboard that improves patient care by enabling physicians to focus on the relevant information. Furthermore, it has been demonstrated that the aspects most often mentioned are not context-specific and can be generalized to other medical contexts.

Unraveling the role of physicochemical differences in predicting protein-protein interactions.

The ability to accurately predict protein-protein interactions is critically important for understanding major cellular processes. However, current experimental and computational approaches for identifying them are technically very challenging and still have limited success. We propose a new computational method for predicting protein-protein interactions using only primary sequence information. It utilizes the concept of physicochemical similarity to determine which interactions will most likely occur. In our approach, the physicochemical features of proteins are extracted using bioinformatics tools for different organisms. Then they are utilized in a machine-learning method to identify successful protein-protein interactions via correlation analysis. It was found that the most important property that correlates most with the protein-protein interactions for all studied organisms is dipeptide amino acid composition (the frequency of specific amino acid pairs in a protein sequence). While current approaches often overlook the specificity of protein-protein interactions with different organisms, our method yields context-specific features that determine protein-protein interactions. The analysis is specifically applied to the bacterial two-component system that includes histidine kinase and transcriptional response regulators, as well as to the barnase-barstar complex, demonstrating the method's versatility across different biological systems. Our approach can be applied to predict protein-protein interactions in any biological system, providing an important tool for investigating complex biological processes' mechanisms.

A multisystem model for understanding stress and adaptation in ethnically and racially diverse youth.

Contemporary biological and contextually based theoretical frameworks have conceptualized how stress exposure may influence adaptation in youth. However, nearly all of this scholarship neglects the role of specific contextual features and/or biological processes that are involved in ethnic-racial minority youth's responses and adaptation to sociocultural stressors. Drawing on the theoretical principles of the developmental psychopathology framework and contemporary models of stress and adaptation, this article proposes a new multisystem model that explains how multiple levels and systems within and outside of individual youth influence their sociocultural adaptation. We provide empirical evidence to support components of this multisystem model. We propose that research based on our new theoretical framework will capture the sociocultural experiences of ethnic-racial minority youth by centering processes that are relevant to their lived experiences, coping, and adjustment. In doing so, this model will inform psychosocial interventions focused on promoting healthy adaptation among ethnic and racial diverse youth. Finally, we offer recommendations to guide future research on stress and adaptation among ethnic and racial diverse youth, in particular, and developmental psychopathology more broadly.

DEEP SIDE-A DEEP LEARNING FRAMEWORK FOR DRUG SIDE EFFECT PREDICTION

Drug failures due to unforeseen adverse effects at clinical trials pose health risks for the participants and lead to substantial financial losses. Side effect prediction algorithms have the potential to guide the drug design process. LINCS L1000 dataset provides a vast resource of cell line gene expression data perturbed by different drugs and creates a knowledge base for context specific features. The state-of-the-art approach that aims at using context specific information relies on only the high quality experiments in LINCS L1000 and discards a large portion of the experiments. In this study, our goal is to boost the prediction performance by utilizing this data to its full extent. We experiment with 5 deep learning architectures. We find that a multimodal architecture produces the best predictive performance among multi-layer perceptron-based architectures when drug chemical structure (CS), and the full set of drug perturbed gene expression profiles (GEX) are used as modalities. Overall, we observe that the CS is more informative than the GEX. A convolutional neural network-based model that uses only SMILES string representation of the drugs achieves the best results and provides 13:0% macro-AUC and 3:1% micro-AUC improvements over the state-of-the-art. We also show that the model is able to predict side effect-drug pairs that are reported in the literature but was missing in the ground truth side effect dataset.

Integrating the Social Context and Learning Ecologies of Digital Natives in EFL Teaching Approaches

Teachers aiming to develop approaches to learning that facilitate an harmonious and spontaneous relationship with EFL learning for twenty-first century digital natives must ensure that the learning process explores contexts pertinent to the learner. The pedagogical reality is determined by context-specific features on which are grounded methodological choices: the participants’ expectations and assumptions, and the teaching and learning features. Within a context of unprecedented exposure and occasions to use English in daily life, the present paper aims to provide pedagogical proposals which integrate learners’ experiences and their learning ecologies with approaches to EFL teaching.

Scaling Social Innovation: A Cross-Cultural Comparative Study of School-Based Mentoring Interventions

Social innovation is increasingly recognised as a powerful vehicle to address unmet societal needs. Nonetheless, research into how contexts and agency interact to determine outcomes/risks of its scaling appears still limited. This study draws on structuration theory to fill some gaps. By comparing two school-based mentoring interventions, it first shows that comparable external catalysts can trigger diverging ambitions and paths to scale. Second, it finds that certain strategic/agentic choices (e.g. entrepreneurial, political, coalition-building skills; evaluation), combined with specific contextual features, help to achieve more rapidly a larger scale of expansion. Finally, it highlights which risks, in this field, more strongly relate to agentic and/or contextual factors, providing insights for research, policy and practice.

SPARTIN: a Bayesian method for the quantification and characterization of cell type interactions in spatial pathology data.

Introduction: The acquisition of high-resolution digital pathology imaging data has sparked the development of methods to extract context-specific features from such complex data. In the context of cancer, this has led to increased exploration of the tumor microenvironment with respect to the presence and spatial composition of immune cells. Spatial statistical modeling of the immune microenvironment may yield insights into the role played by the immune system in the natural development of cancer as well as downstream therapeutic interventions. Methods: In this paper, we present SPatial Analysis of paRtitioned Tumor-Immune imagiNg (SPARTIN), a Bayesian method for the spatial quantification of immune cell infiltration from pathology images. SPARTIN uses Bayesian point processes to characterize a novel measure of local tumor-immune cell interaction, Cell Type Interaction Probability (CTIP). CTIP allows rigorous incorporation of uncertainty and is highly interpretable, both within and across biopsies, and can be used to assess associations with genomic and clinical features. Results: Through simulations, we show SPARTIN can accurately distinguish various patterns of cellular interactions as compared to existing methods. Using SPARTIN, we characterized the local spatial immune cell infiltration within and across 335 melanoma biopsies and evaluated their association with genomic, phenotypic, and clinical outcomes. We found that CTIP was significantly (negatively) associated with deconvolved immune cell prevalence scores including CD8+ T-Cells and Natural Killer cells. Furthermore, average CTIP scores differed significantly across previously established transcriptomic classes and significantly associated with survival outcomes. Discussion: SPARTIN provides a general framework for investigating spatial cellular interactions in high-resolution digital histopathology imaging data and its associations with patient level characteristics. The results of our analysis have potential implications relevant to both treatment and prognosis in the context of Skin Cutaneous Melanoma. The R-package for SPARTIN is available at https://github.com/bayesrx/SPARTIN along with a visualization tool for the images and results at: https://nateosher.github.io/SPARTIN.

DeepSide: A Deep Learning Approach for Drug Side Effect Prediction.

Drug failures due to unforeseen adverse effects at clinical trials pose health risks for the participants and lead to substantial financial losses. Side effect prediction algorithms have the potential to guide the drug design process. LINCS L1000 dataset provides a vast resource of cell line gene expression data perturbed by different drugs and creates a knowledge base for context specific features. The state-of-the-art approach that aims at using context specific information relies on only the high-quality experiments in LINCS L1000 and discards a large portion of the experiments. In this study, our goal is to boost the prediction performance by utilizing this data to its full extent. We experiment with 5 deep learning architectures. We find that a multi-modal architecture produces the best predictive performance among multi-layer perceptron-based architectures when drug chemical structure (CS), and the full set of drug perturbed gene expression profiles (GEX) are used as modalities. Overall, we observe that the CS is more informative than the GEX. A convolutional neural network-based model that uses only SMILES string representation of the drugs achieves the best results and provides 13.0% macro-AUC and 3.1% micro-AUC improvements over the state-of-the-art. We also show that the model is able to predict side effect-drug pairs that are reported in the literature but was missing in the ground truth side effect dataset. DeepSide is available at http://github.com/OnurUner/DeepSide.

Developing a toolkit for implementing evidence-based guidelines to manage hypertension and diabetes in Cambodia: a descriptive case study

BackgroundIn Cambodia, economic development accompanied by health reforms has led to a rapidly ageing population and an increasing incidence and prevalence of noncommunicable diseases. National strategic plans recognize primary care health centres as the focal points of care for treating and managing chronic conditions, particularly hypertension and type 2 diabetes. However, health centres have limited experience in providing such services. This case study describes the process of developing a toolkit to facilitate the use of evidence-based guidelines to manage hypertension and type 2 diabetes at the health-centre level.MethodsWe developed and revised a preliminary toolkit based on the feedback received from key stakeholders. We gathered feedback through an iterative process of group and one-to-one consultations with representatives of the Ministry of Health, provincial health department, health centres and nongovernmental organizations between April 2019 and March 2021.ResultsA toolkit was developed and organized according to the core tasks required to treat and manage hypertension and type 2 diabetes patients. The main tools included patient identification and treatment cards, risk screening forms, a treatment flowchart, referral forms, and patient education material on risk factors and lifestyle recommendations on diet, exercise, and smoking cessation. The toolkit supplements existing guidelines by incorporating context-specific features, including drug availability and the types of medication and dosage guidelines recommended by the Ministry of Health. Referral forms can be extended to incorporate engagement with community health workers and patient education material adapted to the local context. All tools were translated into Khmer and can be modified as needed based on available resources and arrangements with other institutions.ConclusionsOur study demonstrates how a toolkit can be developed through iterative engagement with relevant stakeholders individually and in groups to support the implementation of evidence-based guidelines. Such toolkits can help strengthen the function and capacity of the primary care system to provide care for noncommunicable diseases, serving as the first step towards developing a more comprehensive and sustainable health system in the context of population ageing and caring for patients with chronic diseases.

The dynseq browser track shows context-specific features at nucleotide resolution.

High-throughput experimental platforms have revolutionized the ability to profile biochemical and functional properties of biological sequences such as DNA, RNA and proteins. By collating several data modalities with customizable tracks rendered using intuitive visualizations, genome browsers enable an interactive and interpretable exploration of diverse types of genome profiling experiments and derived annotations. However, existing genome browser tracks are not well suited for intuitive visualization of high-resolution DNA sequence features such as transcription factor motifs. Typically, motif instances in regulatory DNA sequences are visualized as BED-based annotation tracks, which highlight the genomic coordinates of the motif instances but do not expose their specific sequences. Instead, a genome sequence track needs to be cross-referenced with the BED track to identify sequences of motif hits. Even so, quantitative information about the motif instances such as affinity or conservation as well as differences in base resolution from the consensus motif are not immediately apparent. This makes interpretation slow and challenging. This problem is compounded when analyzing several cellular states and/or molecular readouts (such as ATAC-seq and ChIP–seq) simultaneously, as coordinates of enriched regions (peaks) and the set of active transcription factor motifs vary across cell states.

Tracing coalition changes in knowledge in and engagement with childhood obesity prevention to improve intervention implementation

BackgroundWhile most coalition research focuses on studying the effects of peer relationship structure, this study examines the coevolution of coalition structure and behavior across three communities in the U.S. with the goal of identifying coalition dynamics that impact a childhood obesity prevention intervention. MethodsOver two years (2018–2020), three communities within the U.S. participated in a childhood obesity prevention intervention at different times. This intervention was guided by the Stakeholder-Driven Community Diffusion theory, which describes an empirically testable mechanism for promoting community change. Measures are part of the Stakeholder-driven Community Diffusion (SDCD) survey with demonstrated reliability, which include knowledge of and engagement with childhood obesity prevention and social networks. Data from three coalition-committees and their respective networks were used to build three different stochastic actor-oriented models. These models were used to examine the coevolution of coalition structure with coalition behavior (defined a priori as knowledge of and engagement with obesity prevention) among coalition-committee members and their nominated alters (Network A) and coalition-committee members only (Network B). ResultsOverall, coalitions decrease in size and their structure becomes less dense over time. Both Network A and B show a consistent preference to form and sustain ties with those who have more ties. In Network B, there was a trend for those who have higher knowledge scores to increase their number of ties over time. The same trend appeared in Network A but varied based on their peers’ knowledge in and engagement with childhood obesity prevention. Across models, engagement with childhood obesity prevention research was not a significant driver of changes in either coalition network structure or knowledge.ConclusionsThe trends in coalition Network A and B’s coevolution models may point to context-specific features (e.g., ties among stakeholders) that can be leveraged for better intervention implementation. To that end, examining tie density, average path length, network diameter, and the dynamics of each behavior outcome (i.e., knowledge in and engagement with childhood obesity prevention) may help tailor whole-of-community interventions. Future research should attend to additional behavioral variables (e.g., group efficacy) that can capture other aspects of coalition development and that influence implementation, and to testing the efficacy of network interventions after trends have been identified.

Understanding location choice of Syrian refugees from country to neighbourhood level: Opportunities, restrictions and expectations

This study aimed to delve into distinct strands of the location choice (LC) of Syrian refugees by uncovering its determinants at country, city and neighbourhood levels from the narratives, along with the help of content analysis. Findings providing an aggregate explanation of the Syrian emplacement process revealed that Syrian refugees in Turkey, as active agents in the process, make their destination choices within opportunities and restrictions of different scales through the contribution of other agents, places and processes. Findings emphasizing time- and context-specific features - the transitivity, interrelatedness and multi-dimensionality (legal, economic, socio-cultural and spatial) of determinants of LC - showed that the opportunities and restrictions of LC influence Syrians’ future expectations.

Spaces of the local, spaces of the nation: Intersectional bordering practices in post-Brexit Berlin

This article examines the relationship between bordering practices and processes of situated intersectionality by exploring how British migrants encounter and erect borders as they move through Berlin. Through exploring how research participants conceptualise and orientate themselves towards Berlin’s city spaces and how this relates to transnational and translocal processes of classification, I interrogate how processes of racialisation and classification move across European contexts to manifest within localised spaces. The research explores how these intersections work to minimise, accentuate or transfigure one another as inequalities come into being through urban space by placing feminist intersectional approaches in conversation with border studies. By uniquely focusing on a migrant group infrequently considered in European migration literatures, and often regarded as invisible or unproblematic, we can examine how race, class and gender intersect with nationality and how racialised exclusions from European belonging function through everyday processes. I highlight how classification processes have transnational portability and carry intra-European similarities, yet also assuming context-specific features.

Cambodian Teachers' Perceptions of Online Teaching: During and Beyond the COVID-19 Pandemic

This limited topical life history study aims to gain insights into COVID-19’s impacts on teaching at upper secondary schools through Cambodian teachers’ perceptions of online teaching. It presents teachers’ current challenges and needs as well as future impacts on their teaching practices. Online semi-structured interviews were conducted to collect data from 29 subject teachers and their school directors. This study concluded that the COVID-19 pandemic reduced the quality of teaching and learning due to the limited functions of monitoring students rather than limited digital knowledge and skills. The classroom management is still required although the learning is online. The empirical evidence suggests this effect in science disciplines; especially for calculation-related subjects. However, COVID-19 was viewed as providing secondary education with a great deal for implementing the digital revolution of education 4.0 and created some practical issues for policymakers and implementers. Although the findings largely concur with previous literature on online teaching during the pandemic, they also draw context-specific features of the issue.

A Bounded Formulation for The School Bus Scheduling Problem

This paper proposes a new formulation for the school bus scheduling problem (SBSP), which optimizes school start times and bus operation times to minimize transportation cost. The goal is to minimize the number of buses to serve all bus routes such that each route arrives in a time window before school starts. We show that introducing context-specific features, common in many school districts, can lead to a new time-indexed integer linear programming (ILP) formulation. Based on a strengthened version of the linear relaxation of the ILP, we develop a dependent randomized rounding algorithm that yields near-optimal solutions for large-scale problem instances. The efficient formulation and solution approach enable quick generation of multiple solutions to facilitate strategic planning, which we demonstrate with data from two public school districts in the United States. We also generalize our methodologies to solve a robust version of the SBSP.

Toward a Culture-Informed Conceptualization of Child Agency in a Context Characterized by Political and Military Violence. A Qualitative Exploration throughout Experts’ Voices

Our present study sought to qualitative explore the perceptions of experts on the meaning of children agency in a context characterized by ongoing colonial violence and structural racism. Namely, we explored culture and context-specific features of agency, experts’ perceptions about a decolonized definition, and gaps with the mainstream definition of the construct in Western contexts. The study involved 14 participants (N = 8 women), aged between 32 and 70 years with a mean age of 45 years (SD = 9.72) who came from the Gaza Strip, the West Bank (Ramallah, Bethlehem, Hebron and Jenin) and territories currently part of the State of Israel (Jerusalem and Jaffa). They are all mental health professionals in universities, research centres, hospitals and social welfare services. Secondly, the analysis resulted in a map of five themes representing a culturally oriented Palestinian children’s agency model. A threatening context, alleviating factors, healthy agency, aggravating factors, harmful agency. The Palestinian conceptualization of child agency lies in the multifaceted nature of the construct itself re-declined in a context characterized by multiple levels of complexity- cultural, political, social, economic. Our findings might contribute to creating indicators of Palestinian children’s agentic behaviours and a better operationalization of the construct itself.

HyperspectraI Image Classification With Diverse Region Multiscale Feature Extraction and Spectral Imaging

We propose a dual-branch multi-region gaussian pyramid based multi-scale feature extraction and spectral feature extraction network. We use the PCA algorithm to reduce the dimensionality of the data before feature extraction, but the degree of data downscaling is different for the two-branch network. For the spatial information of hyperspectral images, we use a multi-region piecewise gaussian pyramid downsampling method to generate multi-scale and multi-resolution image data and use an improved resnet network to extract spatial information, so that the network can extract specific contextual features of hyperspectral images. For the spectral information of hyperspectral images, we use the method of imaging spectral information, first reducing the dimension of the spectral data and then expanding the spectral data into an image of <i>N</i>&#x00D7;<i>N</i>. By expanding into an image, the resnet network is introduced to extract spectral information, but the number of network layers is different from that of the resnet network for obtaining spatial information, which can solve the problem of low correct classification rate due to changes in similar spectral data. Finally, the spatial and spectral features after the dual-branch feature extraction network are combined into a fully connected network for classification, and the fusion of the two features can improve the classification accuracy. Our experiments on three commonly used datasets show that the method can improve the accuracy of the classifier.

E-governance using mobile applications: A case study of India during the COVID-19 pandemic

Healthcare initiatives backed by electronic-governance (e-governance) principles have contributed well to the extant literature and practice. Governments and healthcare systems across the world were taken aback by the swamping impact of the COVID-19 pandemic. However, they reacted quickly by developing contact-tracing mobile applications (apps) for creating awareness, providing information about various healthcare initiatives, and helping citizens to use the required information in case of emergency. The major challenge was to develop such e-governance interventions in a short time and ensure their quick adoption among the masses. Hence, it is worthwhile to investigate the factors leading to the adoption of such e-governance initiatives, especially in the context of a widespread pandemic situation. The present study is an attempt to analyze the factors driving the intention to use contact tracing mobile apps launched by governments globally during the COVID-19 pandemic. We have conducted the study in the context of India, where the government launched a community-driven contact tracing mobile app for its citizens during the COVID-19 pandemic in April 2020. The study adopted an empirical approach to test how epistemic value, convenience value, conditional value, functional value, and privacy concerns influenced the intention to use this approach. The study found that intention to use such an app was positively influenced by functional value, which in turn was positively influenced by convenience and conditional values. It suggests that the convenience of using the app, perceived seriousness of the pandemic (i.e., conditional value), and utilitarian benefits (i.e., functional value) of the contact-tracing mobile app enhanced its acceptance. However, its novelty (i.e., epistemic value) and privacy concerns are not significant predictors of intention to use. The study recommends that the government should place more emphasis on improving the functional value which is driven by convenience and context-specific features to push the use of an e-governance initiative during the crisis.

Seq-SetNet: directly exploiting multiple sequence alignment for protein secondary structure prediction.

Accurate prediction of protein structure relies heavily on exploiting multiple sequence alignment (MSA) for residue mutations and correlations as this information specifies protein tertiary structure. The widely used prediction approaches usually transform MSA into inter-mediate models, say position-specific scoring matrix or profile hidden Markov model. These inter-mediate models, however, cannot fully represent residue mutations and correlations carried by MSA; hence, an effective way to directly exploit MSAs is highly desirable. Here, we report a novel sequence set network (called Seq-SetNet) to directly and effectively exploit MSA for protein structure prediction. Seq-SetNet uses an 'encoding and aggregation' strategy that consists of two key elements: (i) an encoding module that takes a component homologue in MSA as input, and encodes residue mutations and correlations into context-specific features for each residue; and (ii) an aggregation module to aggregate the features extracted from all component homologues, which are further transformed into structural properties for residues of the query protein. As Seq-SetNet encodes each homologue protein individually, it could consider both insertions and deletions, as well as long-distance correlations among residues, thus representing more information than the inter-mediate models. Moreover, the encoding module automatically learns effective features and thus avoids manual feature engineering. Using symmetric aggregation functions, Seq-SetNet processes the homologue proteins as a sequence set, making its prediction results invariable to the order of these proteins. On popular benchmark sets, we demonstrated the successful application of Seq-SetNet to predict secondary structure and torsion angles of residues with improved accuracy and efficiency. The code and datasets are available through https://github.com/fusong-ju/Seq-SetNet. Supplementary data are available at Bioinformatics online.