Artificial intelligence-enhanced electrocardiogram analysis for identifying cardiac autonomic neuropathy in patients with diabetes.

AIM: To develop and employ machine learning (ML) algorithms to analyse electrocardiograms (ECGs) for the diagnosis of cardiac autonomic neuropathy (CAN). MATERIALS AND METHODS: We used motif and discord extraction techniques, alongside long short-term memory networks, to analyse 12-lead, 10-s ECG tracings to detect CAN in patients with diabetes. The performance of these methods with the support vector machine classification model was evaluated using 10-fold cross validation with the following metrics: accuracy, precision, recall, F1 score, and area under the receiver-operating characteristic curve (AUC). RESULTS: Among 205 patients (mean age 54 ± 17 years, 54% female), 100 were diagnosed with CAN, including 38 with definite or severe CAN (dsCAN) and 62 with early CAN (eCAN). The best model performance for dsCAN classification was achieved using both motifs and discords, with an accuracy of 0.92, an F1 score of 0.92, a recall at 0.94, a precision of 0.91, and an excellent AUC of 0.93 (95% confidence interval [CI] 0.91-0.94). For the detection of any stage of CAN, the approach combining motifs and discords yielded the best results, with an accuracy of 0.65, F1 score of 0.68, a recall of 0.75, a precision of 0.68, and an AUC of 0.68 (95% CI 0.54-0.81). CONCLUSION: Our study highlights the potential of using ML techniques, particularly motifs and discords, to effectively detect dsCAN in patients with diabetes. This approach could be applied in large-scale screening of CAN, particularly to identify definite/severe CAN where cardiovascular risk factor modification may be initiated.

An interpretable framework for sleep posture change detection and postural inactivity segmentation using wrist kinematics.

Sleep posture and movements offer insights into neurophysiological health and correlate with overall well-being and quality of life. Clinical practices utilise polysomnography for sleep assessment, which is intrusive, performed in unfamiliar environments, and requires trained personnel. While sensor technologies such as actigraphy are less invasive alternatives, concerns about their reliability and precision in clinical practice persist. Moreover, the field lacks a universally accepted algorithm, with methods ranging from raw signal thresholding to data-intensive classification models that may be unfamiliar to medical staff. This paper proposes a comprehensive framework for objectively detecting sleep posture changes and temporally segmenting postural inactivity using clinically relevant joint kinematics, measured by a custom-made wearable sensor. The framework was evaluated on wrist kinematic data from five healthy participants during simulated sleep. Intuitive three-dimensional visualisations of kinematic time series were achieved through dimension reduction-based preprocessing, providing an out-of-the-box framework explainability that may be useful for clinical monitoring and diagnosis. The proposed framework achieved up to 99.2% F1-score and 0.96 Pearson's correlation coefficient for posture detection and inactivity segmentation respectively. This work paves the way for reliable home-based sleep movement analysis, serving patient-centred longitudinal care.

RgnTX: Colocalization analysis of transcriptome elements in the presence of isoform heterogeneity and ambiguity.

Colocalization analysis of genomic region sets has been widely adopted to unveil potential functional interactions between corresponding biological attributes, which often serves as the basis for further investigation. A number of methods have been developed for colocalization analysis of genomic elements. However, none of them explicitly considered the transcriptome heterogeneity and isoform ambiguity, making them less appropriate for analyzing transcriptome elements. Here, we developed RgnTX, an R/Bioconductor tool for the colocalization analysis of transcriptome elements with permutation tests. Different from existing approaches, RgnTX directly takes advantage of transcriptome annotation, and offers high flexibility in the null model to simulate realistic transcriptome-wide background, such as the complex alternative splicing patterns. Importantly, it supports the testing of transcriptome elements without clear isoform association, which is often the real scenario due to technical limitations. Proposed package offers a wide selection of pre-defined functions, easy to be utilized by users for visualizing permutation results, calculating shifted z-scores and conducting multiple hypothesis testing under Benjamini-Hochberg correction. Moreover, with synthetic and real datasets, we show that RgnTX novel testing modes return distinct and more significant results compared to existing genome-based methods. We believe RgnTX should make a useful tool to characterize the randomness of the transcriptome, and for conducting statistical association analysis for genomic region sets within the heterogeneous transcriptome. The package now has been accepted by Bioconductor and is freely available at: https://bioconductor.org/packages/RgnTX.

Triple-kernel gated attention-based multiple instance learning with contrastive learning for medical image analysis.

In machine learning, multiple instance learning is a method evolved from supervised learning algorithms, which defines a "bag" as a collection of multiple examples with a wide range of applications. In this paper, we propose a novel deep multiple instance learning model for medical image analysis, called triple-kernel gated attention-based multiple instance learning with contrastive learning. It can be used to overcome the limitations of the existing multiple instance learning approaches to medical image analysis. Our model consists of four steps. i) Extracting the representations by a simple convolutional neural network using contrastive learning for training. ii) Using three different kernel functions to obtain the importance of each instance from the entire image and forming an attention map. iii) Based on the attention map, aggregating the entire image together by attention-based MIL pooling. iv) Feeding the results into the classifier for prediction. The results on different datasets demonstrate that the proposed model outperforms state-of-the-art methods on binary and weakly supervised classification tasks. It can provide more efficient classification results for various disease models and additional explanatory information.

From deterministic to stochastic: an interpretable stochastic model-free reinforcement learning framework for portfolio optimization.

As a fundamental problem in algorithmic trading, portfolio optimization aims to maximize the cumulative return by continuously investing in various financial derivatives within a given time period. Recent years have witnessed the transformation from traditional machine learning trading algorithms to reinforcement learning algorithms due to their superior nature of sequential decision making. However, the exponential growth of the imperfect and noisy financial data that is supposedly leveraged by the deterministic strategy in reinforcement learning, makes it increasingly challenging for one to continuously obtain a profitable portfolio. Thus, in this work, we first reconstruct several deterministic and stochastic reinforcement algorithms as benchmarks. On this basis, we introduce a risk-aware reward function to balance the risk and return. Importantly, we propose a novel interpretable stochastic reinforcement learning framework which tailors a stochastic policy parameterized by Gaussian Mixtures and a distributional critic realized by quantiles for the problem of portfolio optimization. In our experiment, the proposed algorithm demonstrates its superior performance on U.S. market stocks with a 63.1% annual rate of return while at the same time reducing the market value max drawdown by 10% when back-testing during the stock market crash around March 2020.

Geographic encoding of transcripts enabled high-accuracy and isoform-aware deep learning of RNA methylation.

As the most pervasive epigenetic mark present on mRNA and lncRNA, N6-methyladenosine (m6A) RNA methylation regulates all stages of RNA life in various biological processes and disease mechanisms. Computational methods for deciphering RNA modification have achieved great success in recent years; nevertheless, their potential remains underexploited. One reason for this is that existing models usually consider only the sequence of transcripts, ignoring the various regions (or geography) of transcripts such as 3'UTR and intron, where the epigenetic mark forms and functions. Here, we developed three simple yet powerful encoding schemes for transcripts to capture the submolecular geographic information of RNA, which is largely independent from sequences. We show that m6A prediction models based on geographic information alone can achieve comparable performances to classic sequence-based methods. Importantly, geographic information substantially enhances the accuracy of sequence-based models, enables isoform- and tissue-specific prediction of m6A sites, and improves m6A signal detection from direct RNA sequencing data. The geographic encoding schemes we developed have exhibited strong interpretability, and are applicable to not only m6A but also N1-methyladenosine (m1A), and can serve as a general and effective complement to the widely used sequence encoding schemes in deep learning applications concerning RNA transcripts.

A Novel 3D Unsupervised Domain Adaptation Framework for Cross-Modality Medical Image Segmentation.

We consider the problem of volumetric (3D) unsupervised domain adaptation (UDA) in cross-modality medical image segmentation, aiming to perform segmentation on the unannotated target domain (e.g. MRI) with the help of labeled source domain (e.g. CT). Previous UDA methods in medical image analysis usually suffer from two challenges: 1) they focus on processing and analyzing data at 2D level only, thus missing semantic information from the depth level; 2) one-to-one mapping is adopted during the style-transfer process, leading to insufficient alignment in the target domain. Different from the existing methods, in our work, we conduct a first of its kind investigation on multi-style image translation for complete image alignment to alleviate the domain shift problem, and also introduce 3D segmentation in domain adaptation tasks to maintain semantic consistency at the depth level. In particular, we develop an unsupervised domain adaptation framework incorporating a novel quartet self-attention module to efficiently enhance relationships between widely separated features in spatial regions on a higher dimension, leading to a substantial improvement in segmentation accuracy in the unlabeled target domain. In two challenging cross-modality tasks, specifically brain structures and multi-organ abdominal segmentation, our model is shown to outperform current state-of-the-art methods by a significant margin, demonstrating its potential as a benchmark resource for the biomedical and health informatics research community.

Weakly supervised learning of RNA modifications from low-resolution epitranscriptome data.

MOTIVATION: Increasing evidence suggests that post-transcriptional ribonucleic acid (RNA) modifications regulate essential biomolecular functions and are related to the pathogenesis of various diseases. Precise identification of RNA modification sites is essential for understanding the regulatory mechanisms of RNAs. To date, many computational approaches for predicting RNA modifications have been developed, most of which were based on strong supervision enabled by base-resolution epitranscriptome data. However, high-resolution data may not be available. RESULTS: We propose WeakRM, the first weakly supervised learning framework for predicting RNA modifications from low-resolution epitranscriptome datasets, such as those generated from acRIP-seq and hMeRIP-seq. Evaluations on three independent datasets (corresponding to three different RNA modification types and their respective sequencing technologies) demonstrated the effectiveness of our approach in predicting RNA modifications from low-resolution data. WeakRM outperformed state-of-the-art multi-instance learning methods for genomic sequences, such as WSCNN, which was originally designed for transcription factor binding site prediction. Additionally, our approach captured motifs that are consistent with existing knowledge, and visualization of the predicted modification-containing regions unveiled the potentials of detecting RNA modifications with improved resolution. AVAILABILITY IMPLEMENTATION: The source code for the WeakRM algorithm, along with the datasets used, are freely accessible at: https://github.com/daiyun02211/WeakRM. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Automated Social Text Annotation With Joint Multilabel Attention Networks.

Automated social text annotation is the task of suggesting a set of tags for shared documents on social media platforms. The automated annotation process can reduce users' cognitive overhead in tagging and improve tag management for better search, browsing, and recommendation of documents. It can be formulated as a multilabel classification problem. We propose a novel deep learning-based method for this problem and design an attention-based neural network with semantic-based regularization, which can mimic users' reading and annotation behavior to formulate better document representation, leveraging the semantic relations among labels. The network separately models the title and the content of each document and injects an explicit, title-guided attention mechanism into each sentence. To exploit the correlation among labels, we propose two semantic-based loss regularizers, i.e., similarity and subsumption, which enforce the output of the network to conform to label semantics. The model with the semantic-based loss regularizers is referred to as the joint multilabel attention network (JMAN). We conducted a comprehensive evaluation study and compared JMAN to the state-of-the-art baseline models, using four large, real-world social media data sets. In terms of F1 , JMAN significantly outperformed bidirectional gated recurrent unit (Bi-GRU) relatively by around 12.8%-78.6% and the hierarchical attention network (HAN) by around 3.9%-23.8%. The JMAN model demonstrates advantages in convergence and training speed. Further improvement of performance was observed against latent Dirichlet allocation (LDA) and support vector machine (SVM). When applying the semantic-based loss regularizers, the performance of HAN and Bi-GRU in terms of F1 was also boosted. It is also found that dynamic update of the label semantic matrices (JMANd) has the potential to further improve the performance of JMAN but at the cost of substantial memory and warrants further study.

Management of non-urgent paediatric emergency department attendances by GPs: a retrospective observational study.

BACKGROUND: Non-urgent emergency department (ED) attendances are common among children. Primary care management may not only be more clinically appropriate, but may also improve patient experience and be more cost-effective. AIM: To determine the impact on admissions, waiting times, antibiotic prescribing, and treatment costs of integrating a GP into a paediatric ED. DESIGN AND SETTING: Retrospective cohort study explored non-urgent ED presentations in a paediatric ED in north-west England. METHOD: From 1 October 2015 to 30 September 2017, a GP was situated in the ED from 2.00 pm until 10.00 pm, 7 days a week. All children triaged as 'green' using the Manchester Triage System (non-urgent) were considered to be 'GP appropriate'. In cases of GP non-availability, children considered non-urgent were managed by ED staff. Clinical and operational outcomes, as well as the healthcare costs of children managed by GPs and ED staff across the same timeframe over a 2-year period were compared. RESULTS: Of 115 000 children attending the ED over the study period, a complete set of data were available for 13 099 categorised as 'GP appropriate'; of these, 8404 (64.2%) were managed by GPs and 4695 (35.8%) by ED staff. Median duration of ED stay was 39 min (interquartile range [IQR] 16-108 min) in the GP group and 165 min (IQR 104-222 min) in the ED group (P<0.001). Children in the GP group were less likely to be admitted as inpatients (odds ratio [OR] 0.16; 95% confidence interval [CI] = 0.13 to 0.20) and less likely to wait >4 hours before being admitted or discharged (OR 0.11; 95% CI = 0.08 to 0.13), but were more likely to receive antibiotics (OR 1.42; 95% CI = 1.27 to 1.58). Treatment costs were 18.4% lower in the group managed by the GP (P<0.0001). CONCLUSION: Given the rising demand for children's emergency services, GP in ED care models may improve the management of non-urgent ED presentations. However, further research that incorporates causative study designs is required.

MetaTX: deciphering the distribution of mRNA-related features in the presence of isoform ambiguity, with applications in epitranscriptome analysis.

MOTIVATION: The distribution of biological features strongly indicates their functional relevance. Compared to DNA-related features, deciphering the distribution of mRNA-related features is non-trivial due to the existence of isoform ambiguity and compositional diversity of mRNAs. RESULTS: We propose here a rigorous statistical framework, MetaTX, for deciphering the distribution of mRNA-related features. Through a standardized mRNA model, MetaTX firstly unifies various mRNA transcripts of diverse compositions, and then corrects the isoform ambiguity by incorporating the overall distribution pattern of the features through an EM algorithm. MetaTX was tested on both simulated and real data. Results suggested that MetaTX substantially outperformed existing direct methods on simulated datasets, and that a more informative distribution pattern was produced for all the three datasets tested, which contain N6-Methyladenosine sites generated by different technologies. MetaTX should make a useful tool for studying the distribution and functions of mRNA-related biological features, especially for mRNA modifications such as N6-Methyladenosine. AVAILABILITY AND IMPLEMENTATION: The MetaTX R package is freely available at GitHub: https://github.com/yue-wang-biomath/MetaTX.1.0. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

What matters when managing childhood fever in the emergency department? A discrete-choice experiment comparing the preferences of parents and healthcare professionals in the UK.

BACKGROUND: Fever among children is a leading cause of emergency department (ED) attendance and a diagnostic conundrum; yet robust quantitative evidence regarding the preferences of parents and healthcare providers (HCPs) for managing fever is scarce. OBJECTIVE: To determine parental and HCP preferences for the management of paediatric febrile illness in the ED. SETTING: Ten children's centres and a children's ED in England from June 2018 to January 2019. PARTICIPANTS: 98 parents of children aged 0-11 years, and 99 HCPs took part. METHODS: Nine focus-groups and coin-ranking exercises were conducted with parents, and a discrete-choice experiment (DCE) was conducted with both parents and HCPs, which asked respondents to choose their preferred option of several hypothetical management scenarios for paediatric febrile illness, with differing levels of visit time, out-of-pocket costs, antibiotic prescribing, HCP grade and pain/discomfort from investigations. RESULTS: The mean focus-group size was 4.4 participants (range 3-7), with a mean duration of 27.4 min (range 18-46 min). Response rates to the DCE among parents and HCPs were 94.2% and 98.2%, respectively. Avoiding pain from diagnostics, receiving a faster diagnosis and minimising wait times were major concerns for both parents and HCPs, with parents willing-to-pay £16.89 for every 1 hour reduction in waiting times. Both groups preferred treatment by consultants and nurse practitioners to treatment by doctors in postgraduate training. Parents were willing to trade-off considerable increases in waiting times (24.1 min) to be seen by consultants and to avoid additional pain from diagnostics (45.6 min). Reducing antibiotic prescribing was important to HCPs but not parents. CONCLUSIONS: Both parents and HCPs care strongly about reducing visit time, avoiding pain from invasive investigations and receiving diagnostic insights faster when managing paediatric febrile illness. As such, overdue advances in diagnostic capabilities should improve child and carer experience and HCP satisfaction considerably in managing paediatric febrile illness.

Deep-Channel uses deep neural networks to detect single-molecule events from patch-clamp data.

Single-molecule research techniques such as patch-clamp electrophysiology deliver unique biological insight by capturing the movement of individual proteins in real time, unobscured by whole-cell ensemble averaging. The critical first step in analysis is event detection, so called "idealisation", where noisy raw data are turned into discrete records of protein movement. To date there have been practical limitations in patch-clamp data idealisation; high quality idealisation is typically laborious and becomes infeasible and subjective with complex biological data containing many distinct native single-ion channel proteins gating simultaneously. Here, we show a deep learning model based on convolutional neural networks and long short-term memory architecture can automatically idealise complex single molecule activity more accurately and faster than traditional methods. There are no parameters to set; baseline, channel amplitude or numbers of channels for example. We believe this approach could revolutionise the unsupervised automatic detection of single-molecule transition events in the future.

The cost of diagnostic uncertainty: a prospective economic analysis of febrile children attending an NHS emergency department.

BACKGROUND: Paediatric fever is a common cause of emergency department (ED) attendance. A lack of prompt and definitive diagnostics makes it difficult to distinguish viral from potentially life-threatening bacterial causes, necessitating a cautious approach. This may result in extended periods of observation, additional radiography, and the precautionary use of antibiotics (ABs) prior to evidence of bacterial foci. This study examines resource use, service costs, and health outcomes. METHODS: We studied an all-year prospective, comprehensive, and representative cohort of 6518 febrile children (aged < 16 years), attending Alder Hey Children's Hospital, an NHS-affiliated paediatric care provider in the North West of England, over a 1-year period. Performing a time-driven and activity-based micro-costing, we estimated the economic impact of managing paediatric febrile illness, with focus on nurse/clinician time, investigations, radiography, and inpatient stay. Using bootstrapped generalised linear modelling (GLM, gamma, log), we identified the patient and healthcare provider characteristics associated with increased resource use, applying retrospective case-note identification to determine rates of potentially avoidable AB prescribing. RESULTS: Infants aged less than 3 months incurred significantly higher resource use than any other age group, at £1000.28 [95% CI £82.39-£2993.37] per child, (p < 0.001), while lesser experienced doctors exhibited 3.2-fold [95% CI 2.0-5.1-fold] higher resource use than consultants (p < 0.001). Approximately 32.4% of febrile children received antibiotics, and 7.1% were diagnosed with bacterial infections. Children with viral illnesses for whom antibiotic prescription was potentially avoidable incurred 9.9-fold [95% CI 6.5-13.2-fold] cost increases compared to those not receiving antibiotics, equal to an additional £1352.10 per child, predominantly resulting from a 53.9-h increase in observation and inpatient stay (57.1 vs. 3.2 h). Bootstrapped GLM suggested that infants aged below 3 months and those prompting a respiratory rate 'red flag', treatment by lesser experienced doctors, and Manchester Triage System (MTS) yellow or higher were statistically significant predictors of higher resource use in 100% of bootstrap simulations. CONCLUSION: The economic impact of diagnostic uncertainty when managing paediatric febrile illness is significant, and the precautionary use of antibiotics is strongly associated with increased costs. The use of ED resources is highest among infants (aged less than 3 months) and those infants managed by lesser experienced doctors, independent of clinical severity. Diagnostic advances which could increase confidence to withhold antibiotics may yield considerable efficiency gains in these groups, where the perceived risks of failing to identify potentially life-threatening bacterial infections are greatest.

Volumetric image classification using homogeneous decomposition and dictionary learning: A study using retinal optical coherence tomography for detecting age-related macular degeneration.

Three-dimensional (3D) (volumetric) diagnostic imaging techniques are indispensable with respect to the diagnosis and management of many medical conditions. However there is a lack of automated diagnosis techniques to facilitate such 3D image analysis (although some support tools do exist). This paper proposes a novel framework for volumetric medical image classification founded on homogeneous decomposition and dictionary learning. In the proposed framework each image (volume) is recursively decomposed until homogeneous regions are arrived at. Each region is represented using a Histogram of Oriented Gradients (HOG) which is transformed into a set of feature vectors. The Gaussian Mixture Model (GMM) is then used to generate a "dictionary" and the Improved Fisher Kernel (IFK) approach is used to encode feature vectors so as to generate a single feature vector for each volume, which can then be fed into a classifier generator. The principal advantage offered by the framework is that it does not require the detection (segmentation) of specific objects within the input data. The nature of the framework is fully described. A wide range of experiments was conducted with which to analyse the operation of the proposed framework and these are also reported fully in the paper. Although the proposed approach is generally applicable to 3D volumetric images, the focus for the work is 3D retinal Optical Coherence Tomography (OCT) images in the context of the diagnosis of Age-related Macular Degeneration (AMD). The results indicate that excellent diagnostic predictions can be produced using the proposed framework.

Automated "disease/no disease" grading of age-related macular degeneration by an image mining approach.

PURPOSE: To describe and evaluate an automated grading system for age-related macular degeneration (AMD) by color fundus photography. METHODS: An automated "disease/no disease" grading system for AMD was developed based on image-mining techniques. First, image preprocessing was performed to normalize color and nonuniform illumination of the fundus images to define a region of interest and to identify and remove pixels belonging to retinal vessels. To represent images for the prediction task, a graph-based image representation using quadtrees was then adopted. Next, a graph-mining technique was applied to the generated graphs to extract relevant features (in the form of frequent subgraphs) from images of both AMD and healthy volunteers. Features of the training data were then fed into a classifier generator for training purposes before employing the trained classifiers to classify new "unseen" images. RESULTS: The algorithm was evaluated on two publically available fundus-image datasets comprising 258 images (160 AMD and 98 normal). Ten-fold cross validation was used. The experiments produced a best specificity of 100% and a best sensitivity of 99.4% with an overall accuracy of 99.6%. Our approach outperformed previous approaches reported in the literature. CONCLUSIONS: This study has demonstrated a proof-of-concept, image-mining technique for automated AMD grading. This technique has the potential to be further developed as an automated grading tool for future whole-scale AMD screening programs.

Analysis of mesenchymal stem cell differentiation in vitro using classification association rule mining.

In this paper, data mining is used to analyze the data on the differentiation of mammalian Mesenchymal Stem Cells (MSCs), aiming at discovering known and hidden rules governing MSC differentiation, following the establishment of a web-based public database containing experimental data on the MSC proliferation and differentiation. To this effect, a web-based public interactive database comprising the key parameters which influence the fate and destiny of mammalian MSCs has been constructed and analyzed using Classification Association Rule Mining (CARM) as a data-mining technique. The results show that the proposed approach is technically feasible and performs well with respect to the accuracy of (classification) prediction. Key rules mined from the constructed MSC database are consistent with experimental observations, indicating the validity of the method developed and the first step in the application of data mining to the study of MSCs.