Independent Test Data Research Articles

Measuring Activities of Daily Living in Stroke Patients with Motion Machine Learning Algorithms: A Pilot Study

Measuring activities of daily living (ADLs) using wearable technologies may offer higher precision and granularity than the current clinical assessments for patients after stroke. This study aimed to develop and determine the accuracy of detecting different ADLs using machine-learning (ML) algorithms and wearable sensors. Eleven post-stroke patients participated in this pilot study at an ADL Simulation Lab across two study visits. We collected blocks of repeated activity (“atomic” activity) performance data to train our ML algorithms during one visit. We evaluated our ML algorithms using independent semi-naturalistic activity data collected at a separate session. We tested Decision Tree, Random Forest, Support Vector Machine (SVM), and eXtreme Gradient Boosting (XGBoost) for model development. XGBoost was the best classification model. We achieved 82% accuracy based on ten ADL tasks. With a model including seven tasks, accuracy improved to 90%. ADL tasks included chopping food, vacuuming, sweeping, spreading jam or butter, folding laundry, eating, brushing teeth, taking off/putting on a shirt, wiping a cupboard, and buttoning a shirt. Results provide preliminary evidence that ADL functioning can be predicted with adequate accuracy using wearable sensors and ML. The use of external validation (independent training and testing data sets) and semi-naturalistic testing data is a major strength of the study and a step closer to the long-term goal of ADL monitoring in real-world settings. Further investigation is needed to improve the ADL prediction accuracy, increase the number of tasks monitored, and test the model outside of a laboratory setting.

Phase identification for dynamic CT enhancements with generative adversarial network.

Dynamic contrast-enhanced computed tomography (CT) is widely used to provide dynamic tissue contrast for diagnostic investigation and vascular identification. However, the phase information of contrast injection is typically recorded manually by technicians, which introduces missing or mislabeling. Hence, imaging-based contrast phase identification is appealing, but challenging, due to large variations among different contrast protocols, vascular dynamics, and metabolism, especially for clinically acquired CT scans. The purpose of this study is to perform imaging-based phase identification for dynamic abdominal CT using a proposed adversarial learning framework across five representative contrast phases. A generative adversarial network (GAN) is proposed as a disentangled representation learning model. To explicitly model different contrast phases, a low dimensional common representation and a class specific code are fused in the hidden layer. Then, the low dimensional features are reconstructed following a discriminator and classifier. 36350 slices of CT scans from 400 subjects are used to evaluate the proposed method with fivefold cross-validation with splits on subjects. Then, 2216 slices images from 20 independent subjects are employed as independent testing data, which are evaluated using multiclass normalized confusion matrix. The proposed network significantly improved correspondence (0.93) over VGG, ResNet50, StarGAN, and 3DSE with accuracy scores 0.59, 0.62, 0.72, and 0.90, respectively (P<0.001 Stuart-Maxwell test for normalized multiclass confusion matrix). We show that adversarial learning for discriminator can be benefit for capturing contrast information among phases. The proposed discriminator from the disentangled network achieves promising results.

LightCUD: a program for diagnosing IBD based on human gut microbiome data

BackgroundThe diagnosis of inflammatory bowel disease (IBD) and discrimination between the types of IBD are clinically important. IBD is associated with marked changes in the intestinal microbiota. Advances in next-generation sequencing (NGS) technology and the improved hospital bioinformatics analysis ability motivated us to develop a diagnostic method based on the gut microbiome.ResultsUsing a set of whole-genome sequencing (WGS) data from 349 human gut microbiota samples with two types of IBD and healthy controls, we assembled and aligned WGS short reads to obtain feature profiles of strains and genera. The genus and strain profiles were used for the 16S-based and WGS-based diagnostic modules construction respectively. We designed a novel feature selection procedure to select those case-specific features. With these features, we built discrimination models using different machine learning algorithms. The machine learning algorithm LightGBM outperformed other algorithms in this study and thus was chosen as the core algorithm. Specially, we identified two small sets of biomarkers (strains) separately for the WGS-based health vs IBD module and ulcerative colitis vs Crohn’s disease module, which contributed to the optimization of model performance during pre-training.We released LightCUD as an IBD diagnostic program built with LightGBM. The high performance has been validated through five-fold cross-validation and using an independent test data set. LightCUD was implemented in Python and packaged free for installation with customized databases. With WGS data or 16S rRNA sequencing data of gut microbiome samples as the input, LightCUD can discriminate IBD from healthy controls with high accuracy and further identify the specific type of IBD. The executable program LightCUD was released in open source with instructions at the webpage http://cqb.pku.edu.cn/ZhuLab/LightCUD/. The identified strain biomarkers could be used to study the critical factors for disease development and recommend treatments regarding changes in the gut microbial community.ConclusionsAs the first released human gut microbiome-based IBD diagnostic tool, LightCUD demonstrates a high-performance for both WGS and 16S sequencing data. The strains that either identify healthy controls from IBD patients or distinguish the specific type of IBD are expected to be clinically important to serve as biomarkers.

Radiomics signature of brain metastasis: prediction of EGFR mutation status.

To predict epidermal growth factor receptor (EGFR) mutation status in lung adenocarcinoma using MR-based radiomics signature of brain metastasis and explore the optimal MR sequence for prediction. Data from 52 patients with brain metastasis from lung adenocarcinoma (28 with mutant EGFR, 24 with wild-type EGFR) were retrospectively reviewed. Contrast-enhanced T1-weighted imaging (T1-CE), T2 fluid-attenuated inversion recovery (T2-FLAIR), T2WI, and DWI sequences were selected for radiomics features extraction. A total of 438 radiomics features were extracted from each MR sequence. All sequences were randomly divided into training and validation cohorts. The least absolute shrinkage selection operator was used to select informative features, a radiomics signature was built with the logistic regression model of the training cohort, and the radiomics signature performance was evaluated using the validation cohort and an independent testing data set. The radiomics signature built on 9 selected features showed good discrimination in both the training and validation cohorts for T2-FLAIR. The radiomics signature of T2-FLAIR yielded an AUC of 0.987, a classification accuracy of 0.991, sensitivity of 1.000, and specificity of 0.980 in the validation cohort. The AUC was 0.871 in the independent testing data set. The AUCs of our radiomics signature to differentiate exon 19 and exon 21 mutations were 0.529, 0.580, 0.645, and 0.406 for T1-CE, T2-FLAIR, T2WI, and DWI, respectively. We developed a T2-FLAIR radiomics signature that can be used as a noninvasive auxiliary tool for predicting EGFR mutation status in lung adenocarcinoma, which is helpful to guide therapeutic strategies. • MR-based radiomics signature of brain metastasis may help predict EGFR mutation status in lung adenocarcinoma, especially using T2-FLAIR. • Nine radiomics features extracted from T2-FLAIR sequence strongly correlate with EGFR mutation status. • Radiomics features reflect tumor heterogeneity through potential changes in tissue morphology caused by EGFR mutation.

BBPpred: Sequence-Based Prediction of Blood-Brain Barrier Peptides with Feature Representation Learning and Logistic Regression.

Blood-brain barrier peptides (BBPs) have a large range of biomedical applications since they can cross the blood-brain barrier based on different mechanisms. As experimental methods for the identification of BBPs are laborious and expensive, computational approaches are necessary to be developed for predicting BBPs. In this work, we describe a computational method, BBPpred (blood-brain barrier peptides prediction), that can efficiently identify BBPs using logistic regression. We investigate a wide variety of features from amino acid sequence information, and then a feature learning method is adopted to represent the informative features. To improve the prediction performance, seven informative features are selected for classification by eliminating redundant and irrelevant features. In addition, we specifically create two benchmark data sets (training and independent test), which contain a total of 119 BBPs from public databases and the literature. On the training data set, BBPpred shows promising performances with an AUC score of 0.8764 and an AUPR score of 0.8757 using the 10-fold cross-validation. We also test our new method on the independent test data set and obtain a favorable performance. We envision that BBPpred will be a useful tool for identifying, annotating, and characterizing BBPs. BBPpred is freely available at http://BBPpred.xialab.info.

IRC-Fuse: improved and robust prediction of redox-sensitive cysteine by fusing of multiple feature representations.

Redox-sensitive cysteine (RSC) thiol contributes to many biological processes. The identification of RSC plays an important role in clarifying some mechanisms of redox-sensitive factors; nonetheless, experimental investigation of RSCs is expensive and time-consuming. The computational approaches that quickly and accurately identify candidate RSCs using the sequence information are urgently needed. Herein, an improved and robust computational predictor named IRC-Fuse was developed to identify the RSC by fusing of multiple feature representations. To enhance the performance of our model, we integrated the probability scores evaluated by the random forest models implementing different encoding schemes. Cross-validation results exhibited that the IRC-Fuse achieved accuracy and AUC of 0.741 and 0.807, respectively. The IRC-Fuse outperformed exiting methods with improvement of 10% and 13% on accuracy and MCC, respectively, over independent test data. Comparative analysis suggested that the IRC-Fuse was more effective and promising than the existing predictors. For the convenience of experimental scientists, the IRC-Fuse online web server was implemented and publicly accessible at http://kurata14.bio.kyutech.ac.jp/IRC-Fuse/ .

TargetMM: Accurate Missense Mutation Prediction by Utilizing Local and Global Sequence Information with Classifier Ensemble.

Missense mutation (MM) may lead to various human diseases by disabling proteins. Accurate prediction of MM is important and challenging for both protein function annotation and drug design. Although several computational methods yielded acceptable success rates, there is still room for further enhancing the prediction performance of MM. In the present study, we designed a new feature extracting method, which considers the impact degree of residues in the microenvironment range to the mutation site. Stringent cross-validation and independent test on benchmark datasets were performed to evaluate the efficacy of the proposed feature extracting method. Furthermore, three heterogeneous prediction models were trained and then ensembled for the final prediction. By combining the feature representation method and classifier ensemble technique, we reported a novel MM predictor called TargetMM for identifying the pathogenic mutations from the neutral ones. Comparison outcomes based on statistical evaluation demonstrate that TargetMM outperforms the prior advanced methods on the independent test data. The source codes and benchmark datasets of TargetMM are freely available at https://github.com/sera616/TargetMM.git for academic use.

Feature Selection Methods for Protein Biomarker Discovery from Proteomics or Multiomics Data

Untargeted mass spectrometry (MS)-based proteomics provides a powerful platform for protein biomarker discovery, but clinical translation depends on the selection of a small number of proteins for downstream verification and validation. Due to the small sample size of typical discovery studies, protein markers identified from discovery data may not be generalizable to independent datasets. In addition, a good protein marker identified using a discovery platform may be difficult to implement in verification and validation platforms. Moreover, although multiomics characterization is being increasingly used in discovery cohort studies, there is no existing method for multiomics-facilitated protein biomarker selection. Here, we present ProMS, a computational algorithm for protein marker selection. The algorithm is based on the hypothesis that a phenotype is characterized by a few underlying biological functions, each manifested by a group of coexpressed proteins. A weighted k-medoids clustering algorithm is applied to all univariately informative proteins to identify both coexpressed protein clusters and a representative protein for each cluster as markers. In two clinically important classification problems, ProMS shows superior performance compared with existing feature selection methods. ProMS can be extended to the multiomics setting (ProMS_mo) through a constrained weighted k-medoids clustering algorithm, and the protein panels selected by ProMS_mo show improved performance on independent test data compared with ProMS. In addition to superior performance, ProMS and ProMS_mo also have two unique strengths. First, the feature clusters enable functional interpretation of the selected protein markers. Second, the feature clusters provide an opportunity to select replacement protein markers, facilitating a robust transition to the verification and validation platforms. In summary, this study provides a unified and effective computational framework for selecting protein biomarkers using proteomics or multiomics data. The software implementation is publicly available at https://github.com/bzhanglab/proms.

Double data piling leads to perfect classification

Data piling refers to the phenomenon that training data vectors from each class project to a single point for classification. While this interesting phenomenon has been a key to understanding many distinctive properties of high-dimensional discrimination, the theoretical underpinning of data piling is far from properly established. In this work, high-dimensional asymptotics of data piling is investigated under a spiked covariance model, which reveals its close connection to the well-known ridged linear classifier. In particular, by projecting the ridge discriminant vector onto the subspace spanned by the leading sample principal component directions and the maximal data piling vector, we show that a negatively ridged discriminant vector can asymptotically achieve data piling of independent test data, essentially yielding a perfect classification. The second data piling direction is obtained purely from training data and shown to have a maximal property. Furthermore, asymptotic perfect classification occurs only along the second data piling direction.

Deep-learning seismic full-waveform inversion for realistic structural models

Velocity model inversion is one of the most important tasks in seismic exploration. Full-waveform inversion (FWI) can obtain the highest resolution in traditional velocity inversion methods, but it heavily depends on initial models and is computationally expensive. In recent years, a large number of deep-learning (DL)-based velocity model inversion methods have been proposed. One critical component in those DL-based methods is a large training set containing different velocity models. We have developed a method to construct a realistic structural model for the DL network. Our compressional-wave velocity model building method for creating dense-layer/fault/salt body models can automatically construct a large number of models without much human effort, which is very meaningful for DL networks. Moreover, to improve the inversion result on these realistic structural models, instead of only using the common-shot gather, we also extract features from the common-receiver gather as well. Through a large number of realistic structural models, reasonable data acquisition methods, and appropriate network setups, a more generalized result can be obtained through our proposed inversion framework, which has been demonstrated to be effective on the independent testing data set. The results of dense-layer models, fault models, and salt body models that we compared and analyzed demonstrate the reliability of our method and also provide practical guidelines for choosing optimal inversion strategies in realistic situations.

Reliable and Interpretable Mortality Prediction with Strong Foresight in COVID-19 Patients: An International Study from China and Germany

Background: Cohort-independent robust mortality prediction model in patients with COVID-19 infection is not yet established, while this mortality prediction model could help for early triaging, optimizing the medical resource and adjust the treatments in time. Methods: To build up a reliable, interpretable mortality prediction model with strong foresight, we have performed an international, bi-institutional study from China (Wuhan cohort, collected from January to March) and Germany (Würzburg cohort, collected from March to September). A Random Forest-based machine learning approach was applied to 1,352 patients from Wuhan cohort, generating a mortality prediction model based on their clinical features. Then, the 81 patients from Würzburg cohort were used as independent test data for this model. Findings: We identified five clinical features, including lymphocyte (%), neutrophil count, C-reactive protein, lactate dehydrogenase and α-hydroxybutyrate dehydrogenase, could be used for mortality prediction of COVID-19 patients with more than 91% accuracy and 99% AUC. Additionally, the time-series analysis revealed the predictive model based on these clinical features is very robust over time when patients are in the hospital, indicating the strong association of these five clinical features with the progression of treatment as well. Moreover, for different preexisting diseases, this model also demonstrated high predictive power. Finally, the mortality prediction model has been applied to the independent Würzburg cohort, resulting in high prediction accuracy (with above 90% accuracy and 85% AUC) as well, indicating the robustness of the model in different cohorts. Interpretation: This study has established the mortality prediction model that allowed early classification of COVID-19 patients, not only at admission but also along the treatment timeline, not only cohort-independent but also highly interpretable. This model represents a valuable tool for triaging and optimizing the resources in COVID-19 patients. Funding: National Science Foundation of China, National Key Research and Development Program of China and Urgent projects of scientific and technological research on COVID-19 funded by Hubei province.Declaration of Interests: The authors declare no competing interests.Ethics Approval Statement: This study was approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology. Due to the retrospective nature of this study, the local institutional review board of the University of Würzburg waived the requirement for additional approval.

What have biological records ever done for us? A systematic scoping review

Biological records provide biodiversity information over large spatial and temporal scales. Our systematic scoping review of biological records from the well-recorded region of the United Kingdom (UK) and Ireland revealed that over half of all studies using biological records were studying species distributions (134 of 253 studies) and/or temporal trends (139 of 253 studies). A minority of studies (61 of 253) focused on methodological questions, while most studies used biological records with existing methods as tools for answering biological and ecological questions. However, only 31 of 253 studies tested models using independent data. Most studies (154 of 253) integrated multiple biological records datasets, showing that biological records hold a largely untapped potential for independently testing conclusions by withholding some of those datasets for use as independent test data. Our results provide guidance for data providers and researchers interested in more effectively collecting and using biological records.

Quantitative evaluation of liver fibrosis based on ultrasound radio frequency signals: An animal experimental study

BackgroundChronic liver disease is an important cause of liver failure and death worldwide, and liver fibrosis is a common pathological process of most chronic liver diseases. There still lacks a useful tool for evaluating liver fibrosis progression precisely and non-invasively. The purpose of this study was to explore the use of ultrasound radio frequency (RF) signals combined with deep learning approach to evaluate the degree of liver fibrosis quantitatively. MethodsIn this study, by extracting the output of deep learning models as a prediction value, a quantitative liver fibrosis prediction method was achieved based on the bidirectional long short-term memory (Bi-LSTM) network to analyze radio frequency (RF) signals. The dataset consisted of 160 sets of ultrasound RF signals of rat livers, including five fibrosis stages 0−4, upon pathological diagnosis. In total, 150 sets of RF signals were used to train four deep learning classification models, the output of which contained quantitative information. In each training stage of the four models, a large number of signal segments were extracted from the 150 sets and divided randomly into training and validation sets in a ratio of 80:20. Ten sets of RF data using the gold standard of quantitative fibrosis parameter (q-FP) of liver tissues were left for independent testing. To validate the proposed method, correlation analysis was carried out between q-FP and the quantitative prediction results based on the independent test data. ResultsThe accuracy of the four deep learning networks using the training and validation data was above 0.83 and 0.80, and the corresponding areas under the receiver operating characteristic curves were higher than 0.95 and 0.93, respectively. For the quantitative analysis in the independent test set, the determination coefficient, R2, of the linear regression analysis between the quantitative prediction results and q-FP was above 0.93. liver fibrosis is a common pathological process of most chronic liver diseases. ConclusionsThis study indicates that a prediction system based on ultrasound RF signals and a deep learning approach is promising for realizing quantitative and visualized diagnosis of liver fibrosis, which would be of great value in monitoring liver fibrosis non-invasively.

Abstract 16761: Refining Individualized Risk Prediction for Apparent Treatment-Resistant Hypertension in a Large Multiethnic Biobank: The Million Veteran Program

Introduction: Approximately 10-20% of treated hypertensive patients have apparent treatment-resistant hypertension (ATRH), which is associated with adverse cardiovascular outcomes. Our goal was to develop a prediction model for ATRH that combines clinical data and a genetic risk score (GRS) for ATRH. Methods: Study participants were 232,925 patients with hypertension (HTN) enrolled in the Million Veteran Program receiving care at the VA from 2004 to 2017. 21% were black and 6% were women. Incident ATRH was defined as failure to achieve outpatient blood pressure (BP) &lt;140/90 mmHg with 3 anti-HTN drugs including a thiazide or the use of ≥4 drugs. In the training data (n = 207,375), a logistic regression model was fit with age, sex, race, smoking history, body mass index, diabetes, anti-HTN drug classes, baseline systolic BP (SBP), diastolic BP, eGFR, serum potassium, serum bicarbonate and serum lipids. A weighted GRS was developed in the training data using 21 significant SNPs from a transethnic ATRH GWAS. An independent test data (n = 25,550) was used to validate the clinical model and fit a GRS-enriched (clinical model + GRS) model. Results: Over a 13.5-year follow-up, 20,536 and 2,247 incident ATRH cases were observed in the training and test data. In the training data, the clinical model had reasonable discriminant ability (C=0.706) and good calibration (Brier score = 0.085). Age, diabetes, and SBP were the strongest predictors of ATRH. In the test dataset, a 1-IQR increase in the GRS was associated with a 10% (95%CI: 1.04-1.17) higher odds of ATRH. In the test dataset, the LR χ 2 , C-statistic and Brier score for the clinical model were 1075, 0.705 and 0.077 respectively. In the GRS-enriched model, these values were 1107, 0.708 and 0.077. The net reclassification improvement was 1.5% (95% CI: 0.3-2.8, p=0.02). The GRS-enriched model had excellent calibration with an optimism-corrected calibration slope of 0.95. Conclusion: Augmenting a clinical prediction model for ATRH with a GRS comprising GWAS-significant SNPs improved model fit but not discriminant ability. Future work will evaluate whether models incorporating polygenic risk scores with a larger number of variants may refine ATRH risk prediction and stratification and evaluate race-specific performance.

Fully automated contrast and non-contrast cardiac view detection in echocardiography a multi-centre, multi-vendor study

Abstract Background Correctly identifying views acquired in a 2D echocardiographic examination is paramount to post-processing and quantification steps often performed as part of most clinical workflows. In many exams, particularly in stress echocardiography, microbubble contrast is used which greatly affects the appearance of the cardiac views. Here we present a bespoke, fully automated convolutional neural network (CNN) which identifies apical 2, 3, and 4 chamber, and short axis (SAX) views acquired with and without contrast. The CNN was tested in a completely independent, external dataset with the data acquired in a different country than that used to train the neural network. Methods Training data comprised of 2D echocardiograms was taken from 1014 subjects from a prospective multisite, multi-vendor, UK trial with the number of frames in each view greater than 17,500. Prior to view classification model training, images were processed using standard techniques to ensure homogenous and normalised image inputs to the training pipeline. A bespoke CNN was built using the minimum number of convolutional layers required with batch normalisation, and including dropout for reducing overfitting. Before processing, the data was split into 90% for model training (211,958 frames), and 10% used as a validation dataset (23,946 frames). Image frames from different subjects were separated out entirely amongst the training and validation datasets. Further, a separate trial dataset of 240 studies acquired in the USA was used as an independent test dataset (39,401 frames). Results Figure 1 shows the confusion matrices for both validation data (left) and independent test data (right), with an overall accuracy of 96% and 95% for the validation and test datasets respectively. The accuracy for the non-contrast cardiac views of &amp;gt;99% exceeds that seen in other works. The combined datasets included images acquired across ultrasound manufacturers and models from 12 clinical sites. Conclusion We have developed a CNN capable of automatically accurately identifying all relevant cardiac views used in “real world” echo exams, including views acquired with contrast. Use of the CNN in a routine clinical workflow could improve efficiency of quantification steps performed after image acquisition. This was tested on an independent dataset acquired in a different country to that used to train the model and was found to perform similarly thus indicating the generalisability of the model. Figure 1. Confusion matrices Funding Acknowledgement Type of funding source: Private company. Main funding source(s): Ultromics Ltd.

Atrial Fibrillation Detection During Sepsis: Study on MIMIC III ICU Data.

Sepsis is defined by life-threatening organ dysfunction during infection and is one of the leading causes of critical illness. During sepsis, there is high risk that new-onset of atrial fibrillation (AF) can occur, which is associated with significant morbidity and mortality. As a result, computer aided automated and reliable detection of new-onset AF during sepsis is crucial, especially for the critically ill patients in the intensive care unit (ICU). In this paper, a novel automated and robust two-step algorithm to detect AF from ICU patients using electrocardiogram (ECG) signals is presented. First, several statistical parameters including root mean square of successive differences, Shannon entropy, and sample entropy were calculated from the heart rate for the screening of possible AF segments. Next, Poincaré plot-based features along with P-wave characteristics were used to reduce false positive detection of AF, caused by the premature atrial and ventricular beats. A subset of the Medical Information Mart for Intensive Care (MIMIC) III database containing 198 subjects was used in this study. During the training and validation phases, both the simple thresholding as well as machine learning classifiers achieved very high segment-wise AF classification performance. Finally, we tested the performance of our proposed algorithm using two independent test data sets and compared the performance with two state-of-the-art methods. The algorithm achieved an overall 100% sensitivity, 98% specificity, 98.99% accuracy, 98% positive predictive value, and 100% negative predictive value on the subject-wise AF detection, thus showing the efficacy of our proposed algorithm in critically ill sepsis patients. The annotations of the data have been made publicly available for other investigators.

VirPreNet: a weighted ensemble convolutional neural network for the virulence prediction of influenza A virus using all eight segments.

Influenza viruses are persistently threatening public health, causing annual epidemics and sporadic pandemics. The evolution of influenza viruses remains to be the main obstacle in the effectiveness of antiviral treatments due to rapid mutations. Previous work has been investigated to reveal the determinants of virulence of the influenza A virus. To further facilitate flu surveillance, explicit detection of influenza virulence is crucial to protect public health from potential future pandemics. In this article, we propose a weighted ensemble convolutional neural network (CNN) for the virulence prediction of influenza A viruses named VirPreNet that uses all eight segments. Firstly, mouse lethal dose 50 is exerted to label the virulence of infections into two classes, namely avirulent and virulent. A numerical representation of amino acids named ProtVec is applied to the eight-segments in a distributed manner to encode the biological sequences. After splittings and embeddings of influenza strains, the ensemble CNN is constructed as the base model on the influenza dataset of each segment, which serves as the VirPreNet's main part. Followed by a linear layer, the initial predictive outcomes are integrated and assigned with different weights for the final prediction. The experimental results on the collected influenza dataset indicate that VirPreNet achieves state-of-the-art performance combining ProtVec with our proposed architecture. It outperforms baseline methods on the independent testing data. Moreover, our proposed model reveals the importance of PB2 and HA segments on the virulence prediction. We believe that our model may provide new insights into the investigation of influenza virulence. Codes and data to generate the VirPreNet are publicly available at https://github.com/Rayin-saber/VirPreNet. Supplementary data are available at Bioinformatics online.

Radiomic Immunophenotyping of GSEA-Assessed Immunophenotypes of Glioblastoma and Its Implications for Prognosis: A Feasibility Study.

Simple SummaryCharacterization of immunophenotypes in GBM is important for therapeutic stratification and helps predict treatment response and prognosis. However, identifying immunophenotypes of patients with GBM requires multiple laboratory experiments and is time consuming. We developed a non-invasive method to evaluate enrichment levels of CTL, aDC, Treg, and MDSC immune cells to classify immunophenotypes of GBM tumor microenvironment with radiomic features of MR imaging. Five immunophenotypes (G1–G5) of GBM can be classified with specific gene set enrichment analysis. G2 had the worst prognosis and comprised highly enriched MDSCs and lowly enriched CTLs. G3 had the best prognosis and comprised lowly enriched MDSCs and Tregs and highly enriched CTLs. Moreover, the developed radiomics models can successfully identified these two groups by immune cell subsets enriched levels prediction. Therefore, it is possible to characterize immunophenotypes of GBM and predict patient prognosis with radiomics methods.Characterization of immunophenotypes in glioblastoma (GBM) is important for therapeutic stratification and helps predict treatment response and prognosis. Radiomics can be used to predict molecular subtypes and gene expression levels. However, whether radiomics aids immunophenotyping prediction is still unknown. In this study, to classify immunophenotypes in patients with GBM, we developed machine learning-based magnetic resonance (MR) radiomic models to evaluate the enrichment levels of four immune subsets: Cytotoxic T lymphocytes (CTLs), activated dendritic cells, regulatory T cells (Tregs), and myeloid-derived suppressor cells (MDSCs). Independent testing data and the leave-one-out cross-validation method were used to evaluate model effectiveness and model performance, respectively. We identified five immunophenotypes (G1 to G5) based on the enrichment level for the four immune subsets. G2 had the worst prognosis and comprised highly enriched MDSCs and lowly enriched CTLs. G3 had the best prognosis and comprised lowly enriched MDSCs and Tregs and highly enriched CTLs. The average accuracy of T1-weighted contrasted MR radiomics models of the enrichment level for the four immune subsets reached 79% and predicted G2, G3, and the “immune-cold” phenotype (G1) according to our radiomics models. Our radiomic immunophenotyping models feasibly characterize the immunophenotypes of GBM and can predict patient prognosis.

Premature Atrial and Ventricular Contraction Detection using Photoplethysmographic Data from a Smartwatch.

We developed an algorithm to detect premature atrial contraction (PAC) and premature ventricular contraction (PVC) using photoplethysmographic (PPG) data acquired from a smartwatch. Our PAC/PVC detection algorithm is composed of a sequence of algorithms that are combined to discriminate various arrhythmias. A novel vector resemblance method is used to enhance the PAC/PVC detection results of the Poincaré plot method. The new PAC/PVC detection algorithm with our automated motion and noise artifact detection approach yielded a sensitivity of 86% for atrial fibrillation (AF) subjects while the overall sensitivity was 67% when normal sinus rhythm (NSR) subjects were also included. The specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy values for the combined data consisting of both NSR and AF subjects were 97%, 81%, 94% and 92%, respectively, for PAC/PVC detection combined with our automated motion and noise artifact detection approach. Moreover, when AF detection was compared with and without PAC/PVC, the sensitivity and specificity increased from 94.55% to 98.18% and from 95.75% to 97.90%, respectively. For additional independent testing data, we used two datasets: a smartwatch PPG dataset that was collected in our ongoing clinical study, and a pulse oximetry PPG dataset from the Medical Information Mart for Intensive Care III database. The PAC/PVC classification results of the independent testing on these two other datasets are all above 92% for sensitivity, specificity, PPV, NPV, and accuracy. The proposed combined approach to detect PAC and PVC can ultimately lead to better accuracy in AF detection. This is one of the first studies involving detection of PAC and PVC using PPG recordings from a smartwatch. The proposed method can potentially be of clinical importance as this enhanced capability can lead to fewer false positive detections of AF, especially for those NSR subjects with frequent episodes of PAC/PVC.

Learning What Makes Catalysts Good

Machine learning has proven a powerful tool for accelerating the computational characterization of energy materials. The neural-network approach by Lu et al. allows bypassing time-consuming first principles calculations for the design of catalysts based on high-entropy alloys. This work is an example of a growing number of case studies identifying descriptors of catalytic performance using machine learning instead of physical intuition.