Similar Predictive Accuracy Research Articles

Dopaminergic Gene Dosage Reveals Distinct Biological Partitions between Autism and Developmental Delay as Revealed by Complex Network Analysis and Machine Learning Approaches.

The neurobiological mechanisms underlying Autism Spectrum Disorders (ASD) remains controversial. One factor contributing to this debate is the phenotypic heterogeneity observed in ASD, which suggests that multiple system disruptions may contribute to diverse patterns of impairment which have been reported between and within study samples. Here, we used SFARI data to address genetic imbalances affecting the dopaminergic system. Using complex network analysis, we investigated the relations between phenotypic profiles, gene dosage and gene ontology (GO) terms related to dopaminergic neurotransmission from a polygenic point-of-view. We observed that the degree of distribution of the networks matched a power-law distribution characterized by the presence of hubs, gene or GO nodes with a large number of interactions. Furthermore, we identified interesting patterns related to subnetworks of genes and GO terms, which suggested applicability to separation of clinical clusters (Developmental Delay (DD) versus ASD). This has the potential to improve our understanding of genetic variability issues and has implications for diagnostic categorization. In ASD, we identified the separability of four key dopaminergic mechanisms disrupted with regard to receptor binding, synaptic physiology and neural differentiation, each belonging to particular subgroups of ASD participants, whereas in DD a more unitary biological pattern was found. Finally, network analysis was fed into a machine learning binary classification framework to differentiate between the diagnosis of ASD and DD. Subsets of 1846 participants were used to train a Random Forest algorithm. Our best classifier achieved, on average, a diagnosis-predicting accuracy of 85.18% (sd 1.11%) on the test samples of 790 participants using 117 genes. The achieved accuracy surpassed results using genetic data and closely matched imaging approaches addressing binary diagnostic classification. Importantly, we observed a similar prediction accuracy when the classifier uses only 62 GO features. This result further corroborates the complex network analysis approach, suggesting that different genetic causes might converge to the dysregulation of the same set of biological mechanisms, leading to a similar disease phenotype. This new biology-driven ontological framework yields a less variable and more compact domain-related set of features with potential mechanistic generalization. The proposed network analysis, allowing for the determination of a clearcut biological distinction between ASD and DD (the latter presenting much lower modularity and heterogeneity), is amenable to machine learning approaches and provides an interesting avenue of research for the future.

Evolving asymmetric and anisotropic hardening of CP-Ti sheets under monotonic and reverse loading: Characterization and modeling

Commercially pure titanium (CP-Ti) sheets are promising to manufacture high-performance and lightweight components in many fields such as aerospace, marine, energy, electrics and healthcare. However, due to its low symmetric hexagonal close-packed (HCP) structure and complex thermal-mechanical processing history, CP-Ti sheets may exhibit significant asymmetry and anisotropy under monotonic and reverse loading. Therefore, the coupling effect of loading-path dependent asymmetric and anisotropic hardening behaviors of CP-Ti sheets need to be systemically characterized and accurately described under monotonic and reverse loading paths. In this study, firstly, the rheological tests were performed to systemically explore the evolving asymmetric and anisotropic hardening behavior of CP-Ti sheet under monotonic and reverse loading modes. With the help of anti-buckling support fixture and DIC system in tension and compression tests, the coupling effect of asymmetric and anisotropic hardening under monotonic and reverse loading of CP-Ti sheet was identified. Secondly, a loading-path dependent constitutive modeling framework was constructed in which the stress invariants-based model (SIM) integrated with distorted hardening (DH) under reverse loading (RL) conditions (SIM+DH+RL) were considered. Within the proposed modeling framework, a judging criterion of stress reversal for explicit solution method and an activation criterion of different reverse loading modes were introduced. Finally, the proposed SIM+DH+RL model was implemented, calibrated and further applied to simulate typical V-/U-bending of CP-Ti sheets. The results show that the proposed SIM+DH+RL model can significantly improve the prediction accuracy of V-/U-bending springback with the relative errors less than 7.8% and 2.1%, compared with 15.2% and 12% relative errors of Mises+IH, 19.5% and 7.3% relative errors of SIM+DH, 15.1% and 8.4% relative errors of Yoon2014+IH, 14.5% and 5.2% relative errors of CPB06+IH, and 17.3% and 11.5% relative errors of Hill48+IH. It is noted that for the thickness distributions after V-/U-bending springback, the above used models have similar prediction accuracies with the relative errors less than 1.8%.

A scalable Bayesian functional GWAS method accounting for multivariate quantitative functional annotations with applications for studying Alzheimer disease.

SummaryExisting methods for integrating functional annotations in genome-wide association studies (GWASs) to fine-map and prioritize potential causal variants are limited to using non-overlapped categorical annotations or limited by the computation burden of modeling genome-wide variants. To overcome these limitations, we propose a scalable Bayesian functional GWAS method to account for multivariate quantitative functional annotations (BFGWAS_QUANT), accompanied by a scalable computation algorithm enabling joint modeling of genome-wide variants. Simulation studies validated the performance of BFGWAS_QUANT for accurately quantifying annotation enrichment and improving GWAS power. Applying BFGWAS_QUANT to study five Alzheimer disease (AD)-related phenotypes using individual-level GWAS data (n = ∼1,000), we found that histone modification annotations have higher enrichment than expression quantitative trait locus (eQTL) annotations for all considered phenotypes, with the highest enrichment in H3K27me3 (polycomb regression). We also found that cis-eQTLs in microglia had higher enrichment than eQTLs of bulk brain frontal cortex tissue for all considered phenotypes. A similar enrichment pattern was also identified using the International Genomics of Alzheimer’s Project (IGAP) summary-level GWAS data of AD (n = ∼54,000). The strongest known APOE E4 risk allele was identified for all five phenotypes, and the APOE locus was validated using the IGAP data. BFGWAS_QUANT fine-mapped 32 significant variants from 1,073 genome-wide significant variants in the IGAP data. We also demonstrated that the polygenic risk scores (PRSs) using effect size estimates by BFGWAS_QUANT had a similar prediction accuracy as other methods assuming a sparse causal model. Overall, BFGWAS_QUANT is a useful GWAS tool for quantifying annotation enrichment and prioritizing potential causal variants.

Prediction of field winter wheat yield using fewer parameters at middle growth stage by linear regression and the BP neural network method

Simple and reliable methods for wheat yield forecasting at the field scale are significant for farmers to formulate effective and timely field management and estimate economic revenue. In this study, the linear regression method (least absolute shrinkage and selection operator regression, LASSO) and BP neural network method (back propagation neural network, BPNN) were chosen for field yield prediction. The four variable types considered are climate, soil moisture, crop growth, and flag leaf photosynthesis. Crop-related data were collected in 12 winter wheat varieties that are widely grown in the North China Plain from 2018 to 2021. The same data set was used for both LASSO and BPNN models’ development and validation. The results show that models based on LASSO method had the determination coefficient (R2) from 0.23 to 0.78, and the root mean square error (RMSE) from 548 to 1023 kg ha−1, while they were 0.66–0.94 and 302–684 kg ha−1 in models based on the BPNN method. This showed that the BPNN method performed better in field wheat yield prediction than the LASSO method. For BPNN method, using only a single variable at the early grain-filling stage had the similar prediction accuracy to using single or comprehensive variables in multiple growth periods, indicating that the field wheat yield can be predicted earlier by BPNN method with fewer indicators. Last, an optimal field wheat prediction model was developed by the BPNN method using crop growth variables (dry matter and leaf area index) at the early grain-filling stage, with the R2 and RMSE of 0.93 and 288 kg ha−1 in the model development stage, respectively, and correspondingly 0.86 and 481 kg ha−1 in the model validation stage. The significant point of this model is that the field wheat yield can be predicted with only two growth indicators approximately 35 days before harvest for 12 mostly cultivated wheat varieties.

330.5: A Prediction Model for Antibody Response to Three Vaccines Against SARS-CoV-2 Among Solid Organ Transplant Recipients

Background: Solid organ transplant recipients (SOTRs) show impaired immune response to SARS-CoV-2 vaccines. A prediction model for immune response to SARS-CoV-2 vaccines could enable targeted antibody monitoring and vaccine dose adjustments in SOTRs during potential epidemic/endemic circulations of SARS-CoV-2. Methods: We studied 782 SOTRs recruited from across the United States who received a 3-dose series of SARS-CoV-2 vaccines and underwent serologic testing for SARS-CoV-2 antibodies between 2-5 weeks after the third dose. We used gradient boosting and logistic regression to create prediction models for positive antibody response based on demographic, transplant-related, and immunosuppression-related clinical factors. Results: Overall, 643 (82.3%) SOTRs had positive antibody responses. The gradient boosting model predicted positive antibody response with moderate accuracy (C-statistic, 0.790). The influential predictors of positive antibody response were mycophenolate avoidance (variable importance, 43.0%), longer time since transplant (31.2%), and corticosteroid avoidance (11.9%). The logistic model showed similar prediction accuracy (C-statistic, 0.790). Mycophenolate use (aOR, 0.07 [95% CI, 0.02-0.16]), corticosteroid use (aOR, 0.39 [95% CI, 0.24-0.61]), and time since transplant (per year; aOR, 1.10 [95% CI, 1.06-1.15]) showed significant association with positive antibody response. Conclusion: Antibody response to a 3-dose SARS-CoV-2 vaccine series could be predicted with moderate accuracy in SOTRs. This prediction model could identify SOTRs with higher risks that warrant additional vaccine doses and antibody monitoring. This work was supported by the Ben-Dov family; grants T32DK007713 (Dr. Alejo), The ASTS Fryer Resident Scientist Award (Dr. Mitchell), K01DK101677 (Dr. Massie), and K23DK115908 (Dr. Garonzik-Wang) from the National Institute of Diabetes and Digestive and Kidney Diseases; grant K24AI144954 (Dr. Segev) and K23AI157893 (Werbel) from the National Institute of Allergy and Infectious Diseases.

Classification model with weighted regularization to improve the reproducibility of neuroimaging signature selection.

Machine learning (ML) has been extensively applied in brain imaging studies to aid the diagnosis of psychiatric disorders and the selection of potential biomarkers. Due to the high dimensionality of imaging data and heterogeneous subtypes of psychiatric disorders, the reproducibility of ML results in brain imaging studies has drawn increasing attention. The reproducibility in brain imaging has been primarily examined in terms of prediction accuracy. However, achieving high prediction accuracy and discovering relevant features are two separate but related goals. An important yet under-investigated problem is the reproducibility of feature selection in brain imaging studies. We propose a new metric to quantify the reproducibility of neuroimaging feature selection via bootstrapping. We estimate the reproducibility index (R-index) for each feature as the reciprocal coefficient of variation of absolute mean difference across a larger number of bootstrap samples. We then integrate the R-index in regularized classification models as penalty weight. Reproducible features with a larger R-index are assigned smaller penalty weights and thus are more likely to be selected by our proposed models. Both simulated and multimodal neuroimaging data are used to examine the performance of our proposed models. Results show that our proposed R-index models are effective in separating informative features from noise features. Additionally, the proposed models yield similar or higher prediction accuracy than the standard regularized classification models while further reducing coefficient estimation error. Improvements achieved by the proposed models are essential to advance our understanding of the selected brain imaging features as well as their associations with psychiatric disorders.

Antibody therapy reverses biological signatures of COVID-19 progression

SummaryUnderstanding who is at risk of progression to severe coronavirus disease 2019 (COVID-19) is key to clinical decision making and effective treatment. We study correlates of disease severity in the COMET-ICE clinical trial that randomized 1:1 to placebo or to sotrovimab, a monoclonal antibody for the treatment of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (ClinicalTrials.gov04545060). Laboratory parameters identify study participants at greater risk of severe disease, including a high neutrophil-to-lymphocyte ratio (NLR), a negative SARS-CoV-2 serologic test, and whole-blood transcriptome profiles. Sotrovimab treatment is associated with normalization of NLR and the transcriptomic profile and with a decrease of viral RNA in nasopharyngeal samples. Transcriptomics provides the most sensitive detection of participants who would go on to be hospitalized or die. To facilitate timely measurement, we identify a 10-gene signature with similar predictive accuracy. We identify markers of risk for disease progression and demonstrate that normalization of these parameters occurs with antibody treatment of established infection.

Physics-based deep learning for modeling nonlinear pulse propagation in optical fibers.

A physics-based deep learning (DL) method termed Phynet is proposed for modeling the nonlinear pulse propagation in optical fibers totally independent of the ground truth. The presented Phynet is a combination of a handcrafted neural network and the nonlinear Schrödinger physics model. In particular, Phynet is optimized through physics loss generated by the interaction between the network and the physical model rather than the supervised loss. The inverse pulse propagation problem is leveraged to exemplify the performance of Phynet when in comparison to the typical DL method under the same structure and datasets. The results demonstrate that Phynet is able to precisely restore the initial pulse profiles with varied initial widths and powers, while revealing a similar prediction accuracy compared with the typical DL method. The proposed Phynet method can be expected to break the severe bottleneck of the traditional DL method in terms of relying on abundant labeled data during the training phase, which thus brings new insight for modeling and predicting the nonlinear dynamics of the fibers.

Predicting shear wave velocity from conventional well logs with deep and hybrid machine learning algorithms

Shear wave velocity (VS) data from sedimentary rock sequences is a prerequisite for implementing most mathematical models of petroleum engineering geomechanics. Extracting such data by analyzing finite reservoir rock cores is very costly and limited. The high cost of sonic dipole advanced wellbore logging service and its implementation in a few wells of a field has placed many limitations on geomechanical modeling. On the other hand, shear wave velocity VS tends to be nonlinearly related to many of its influencing variables, making empirical correlations unreliable for its prediction. Hybrid machine learning (HML) algorithms are well suited to improving predictions of such variables. Recent advances in deep learning (DL) algorithms suggest that they too should be useful for predicting VS for large gas and oil field datasets but this has yet to be verified. In this study, 6622 data records from two wells in the giant Iranian Marun oil field (MN#163 and MN#225) are used to train HML and DL algorithms. 2072 independent data records from another well (MN#179) are used to verify the VS prediction performance based on eight well-log-derived influencing variables. Input variables are standard full-set recorded parameters in conventional oil and gas well logging data available in most older wells. DL predicts VS for the supervised validation subset with a root mean squared error (RMSE) of 0.055 km/s and coefficient of determination (R2) of 0.9729. It achieves similar prediction accuracy when applied to an unseen dataset. By comparing the VS prediction performance results, it is apparent that the DL convolutional neural network model slightly outperforms the HML algorithms tested. Both DL and HLM models substantially outperform five commonly used empirical relationships for calculating VS from Vp relationships when applied to the Marun Field dataset. Concerns regarding the model's integrity and reproducibility were also addressed by evaluating it on data from another well in the field. The findings of this study can lead to the development of knowledge of production patterns and sustainability of oil reservoirs and the prevention of enormous damage related to geomechanics through a better understanding of wellbore instability and casing collapse problems.Graphical abstract

Nutritional quality screening of oat groats by vibrational spectroscopy using field-portable instruments

This study evaluated the performance of low-cost, real-time, and field-deployable spectroscopic instruments operating at near-infrared (NIR) and mid-infrared (MIR) wavelengths for measuring quality traits (β-glucan, starch, protein, and lipid) of oats to support breeding selection. Samples were kindly provided by PepsiCo R&D (n = 150) as oat groats. A handheld FT-NIR sensor (1350–2560 nm) measured spectra of ground and intact oat samples, while a portable FT-IR spectrometer (4000–650 cm −1) measured ground samples only. Several laboratory reference methods were used to measure β-glucan, starch, protein, and lipid composition to develop spectroscopic analysis models based on Partial Least Squares Regression (PLSR). Best model performance was obtained from NIR spectra of ground groats, with standard error of prediction (SEP) for β-glucan, starch, protein, and lipid of 0.2%, 1.0%, 0.6%, and 0.3%, respectively. PLSR models for the MIR spectra exhibited similar predictive accuracy. The performance of these PLSR models either matched or outperformed NIR techniques reported in the literature using portable and benchtop systems. Therefore, novel miniaturized NIR sensors can provide breeders with a rapid method (15 s) to screen for unique traits in the field with equivalent reliability and sensitivity as benchtop systems.

Estimations of the weather effects on brain functions using functional MRI: A cautionary note.

The influences of environmental factors such as weather on the human brain are still largely unknown. A few neuroimaging studies have demonstrated seasonal effects, but were limited by their cross‐sectional design or sample sizes. Most importantly, the stability of the MRI scanner has not been taken into account, which may also be affected by environments. In the current study, we analyzed longitudinal resting‐state functional MRI (fMRI) data from eight individuals, where they were scanned over months to years. We applied machine learning regression to use different resting‐state parameters, including the amplitude of low‐frequency fluctuations (ALFF), regional homogeneity (ReHo), and functional connectivity matrix, to predict different weather and environmental parameters. For careful control, the raw EPI and the anatomical images were also used for predictions. We first found that daylight length and air temperatures could be reliably predicted with cross‐validation using the resting‐state parameters. However, similar prediction accuracies could also be achieved by using one frame of EPI image, and even higher accuracies could be achieved by using the segmented or raw anatomical images. Finally, the signals outside of the brain in the anatomical images and signals in phantom scans could also achieve higher prediction accuracies, suggesting that the predictability may be due to the baseline signals of the MRI scanner. After all, we did not identify detectable influences of weather on brain functions other than the influences on the baseline signals of MRI scanners. The results highlight the difficulty of studying long‐term effects using MRI.

Association between Rapid Dementia Screening Test score and clinical events in elderly patients with cardiovascular disease: a retrospective cohort study.

Cognitive decline is prevalent among patients with cardiovascular disease (CVD). Cognitive measurement has been considered as a standard assessment for secondary prevention; however, standard cognitive tests are sometimes infeasible due to time constraints. This study aimed to examine the association between the Rapid Dementia Screening Test (RDST), a brief screening tool for cognitive function, and clinical events in elderly patients with CVD. This retrospective cohort study included 140 hospitalized patients with CVD who participated in inpatient cardiac rehabilitation (median age, 75 years; male, 67%). Cognitive function for each patient was assessed using the RDST and Montreal Cognitive Assessment (MoCA), a standard test of mild cognitive impairment. The clinical events assessed as outcomes included all-cause mortality and unplanned rehospitalization. Receiver-operating characteristic (ROC) curve analysis showed similar predictive accuracy for the study outcome (P = 0.337) between the RDST [area under the curve, 0.651; 95% confidence interval (CI), 0.559-0.743] and MoCA (0.625; 0.530-0.720). The ROC analysis identified a cut-off value of 9 points for the RDST (sensitivity, 77.8%; specificity, 50.5%). Patients with RDST ≤9 showed a poor survival rate compared with those with ≥10 points (log-rank test, P = 0.002; hazard ratio, 2.94, 95% CI, 1.46-5.94). This result was consistent even after adjusting for potential confounders. The RDST was associated with clinical events in elderly patients with CVD and its predictive capability was comparable with that of MoCA, a standard cognitive test. The RDST may be useful in CVD as an alternative screening tool for cognitive decline.

Comparison of sampling designs for calibrating digital soil maps at multiple depths

Digital soil mapping (DSM) aims to produce detailed maps of soil properties or soil classes to improve agricultural management and soil quality assessment. Optimized sampling design can reduce the substantial costs and efforts associated with sampling, profile description, and laboratory analysis. The purpose of this study was to compare common sampling designs for DSM, including grid sampling (GS), grid random sampling (GRS), stratified random sampling (StRS), and conditioned Latin hypercube sampling (cLHS). In an agricultural field (11 ha) in Quebec, Canada, a total of unique 118 locations were selected using each of the four sampling designs (45 locations each), and additional 30 sample locations were selected as an independent testing dataset (evaluation dataset). Soil visible near-infrared (Vis-NIR) spectra were collected in situ at the 148 locations (1 m depth), and soil cores were collected from a subset of 32 locations and subdivided at 10-cm depth intervals, totaling 251 samples. The Cubist model was used to elucidate the relationship between Vis-NIR spectra and soil properties (soil organic matter (SOM) and clay), which was then used to predict the soil properties at all 148 sample locations. Digital maps of soil properties at multiple depths for the entire field (148 sample locations) were prepared using a quantile random forest model to obtain complete model maps (CM-maps). Soil properties were also mapped using the samples from each of the 45 locations for each sampling design to obtain sampling design maps (SD-maps). The SD-maps were evaluated using the independent testing dataset (30 sample locations), and the spatial distribution and model uncertainty of each SD-map were compared with those of the corresponding CM-map. The spatial and feature space coverage were compared across the four sampling designs. The results showed that GS resulted in the most even spatial coverage, cLHS resulted in the best coverage of the feature space, and GS and cLHS resulted in similar prediction accuracies and spatial distributions of soil properties. The SOM content was underestimated using GRS, with large errors at 0–50 cm depth, due to some values not being captured by this sampling design, whereas larger errors for the deeper soil layers were produced using StRS. Predictions of SOM and clay contents had higher accuracy for topsoil (0–30 cm) than for deep subsoil (60–100 cm). It was concluded that the soil sampling designs with either good spatial coverage or feature space coverage can provide good accuracy in 3D DSM, but their performances may be different for different soil properties.

Evaluation of genomic selection for high salinity tolerance traits in Pacific white shrimp Litopenaeus vannamei

The Pacific white shrimp (Litopenaeus vannamei) can tolerate wide range of salinity from 0.5 to 40. However, the growth and survival rate of shrimp is relatively low when they are cultured in high salinity seawater (>40). There is a large area of high salinity water which could be utilized for shrimp aquaculture. It is necessary to select new varieties of shrimp with high salinity tolerance, which can be cultured in these areas. In this study, the genetic parameters and genomic prediction accuracy on high salinity tolerance traits of Litopenaeus vannamei were evaluated using genome-wide SNPs. The survival time, lethal salinity and survival status of shrimp during high salinity treatments were recorded as phenotypes for each individual. The heritabilities of three phenotypes including survival time, lethal salinity and survival status were 0.38 ± 0.1, 0.40 ± 0.1, 0.5 ± 0.1, respectively. Among different GS models, GBLUP and BayesB showed similar prediction accuracy. The prediction accuracy of GBLUP for survival time, lethal salinity, and survival state of shrimp were 0.62 ± 0.03, 0.64 ± 0.04, and 0.76 ± 0.01, respectively. Compared with traditional pedigree selection (PBLUP), GS models showed a higher prediction accuracy by up to 12.1%. In addition, the optimal training population size was predicted to be around 1000, in which the prediction accuracy could rise to around 0.9. These results indicate that the heritability of high salinity tolerance traits of shrimp is medium to high heritability, and the genomic selection is superior to traditional selection approach. This study provides important guidance for genomic selection of shrimp tolerant to high salinity.

Near-Infrared Spectroscopy to Predict Provitamin A Carotenoids Content in Maize

Vitamin A deficiency (VAD) is a public health issue worldwide. Provitamin A (PVA) biofortified maize serves as an alternative to help combat VAD. Breeding efforts to develop maize varieties with high PVA carotenoid content combine molecular and phenotypic selection strategies. The phenotypic assessment of carotenoids is currently done using liquid chromatography, a precise but time- and resource-consuming methodology. Using near-infrared spectroscopy (NIRS) could increase the breeding efficiency. This study used ultra-performance liquid chromatography (UPLC) data from 1857 tropical maize genotypes as a training set and NIRS data to do an independent test of a set of 650 genotypes to predict PVA carotenoids using Bayesian and modified partial least square (MPLS) regression models. Both regression methods produced similar prediction accuracies for the total carotenoids (r2 = 0.75), lutein (r2 = 0.55), zeaxanthin (r2= 0.61), β-carotene (r2 = 0.22) and β-cryptoxanthin (BCX) (r2 = 0.57). These results demonstrate that Bayesian and MPLS regression of BCX on NIRS data can be used to predict BCX content, the current focus on PVA enhancement, and thus offers opportunities for high-throughput phenotyping at a low cost, especially in the early stages of PVA maize breeding pipeline when many genotypes must be screened.

RNA-Binding Protein COL14A1, TNS1, NUSAP1 and YWHAE Are Valid Biomarkers to Predict Peritoneal Metastasis in Gastric Cancer.

Gastric cancer (GC) is the third leading cause of tumor related mortality worldwide. Peritoneal metastasis (PM) occurs in more than half of advanced GC patients, leading to poor prognosis. Therefore, the GSE62254 cohort was used to construct a signature consisting of four RNA-binding proteins (RBP) to predict the possibility of PM in GC patients. Then, ROC curves were plotted followed by calculation of AUCs, showing that the signature had a similar predictive accuracy compared with the TNM staging system. Importantly, the capability of prediction was enhanced by combining the classifier and TNM staging. In order to validate the expression of the four RBPs in GC tissues with and without PM, immunohistochemistry was further performed on samples from 108 patients. The differential expression of COL14A1, TNS1, NUSAP1 and YWHAE was in accordance with the emergence of PM. Afterwards, we produced Kaplan–Meier curves according to the signature and differential expression of the RBPs in patients. Finally, CCK-8 assays were performed to verify the effect on cell proliferation, finding that COL14A1 and TNS1 promoted cell proliferation, while NUSAP1 and YWHAE led to suppressed cell proliferation. In conclusion, the four-RBP-based signature, combined with TNM staging, has the potential to predict risk of PM in GC.

Detection of Sugarcane Mosaic Diseases Using Deep Learning Architecture to Avoid Annealing Temperature of PCR Primer in Laboratory Testing

The sugarcane leaf diseases such as mosaic and streak mosaic are difficult to differentiate using Image processing techniques because both diseases show similar visual attributes such as pattern and color. To identify the type of diseases, we need to perform Polymerase Chain Reaction (PCR) testing which is used for the classification of diseases in laboratories. The accuracy of the PCR test depends on reaction mix preparation, reaction time, and DNA/RNA extraction. The major problem influencing the PCR test accuracy is the Annealing temperature of the primers and needs a standardized set of samples. In addition, it is a time-consuming process. In this paper, we proposed a Diversified Deep Learning Architecture (DDLA) which is developed with the input images after various pre-processing steps such as denoising using Discrete Wavelet Transform (DWT) and enhancing using histogram equalization in HSI color space to improve the similar pattern disease prediction accuracy. The performance of the proposed model is analyzed for a set of diseased leaves and the results are compared with the output of the popular pertained models such asVGG16, InceptionV3, ResNET50, Inception ResNET, and DenseNET201 with and without pre-processing. The training accuracy of the proposed model is 97% and the testing accuracy is 87%. The DDLA model produces ground truth test results with an accuracy of 88.7% for mosaic and 85.7% for streak mosaic with a less computational time of 152sec compared to the lab test duration of 6 hrs. The performance of the model is also measured in terms of Precision, F1 Score, Specificity, and Sensitivity. The Proposed DDLA model’s F1 score is higher than the pre-trained models with a minimum test loss of 1.167. Moreover, the DDLA structure occupies less memory space when compared to the pertained models.

A Sparse Gating Convolutional Recurrent Network for Traffic Flow Prediction

With the continuous development of deep learning, more and more huge deep learning models are developed by researchers, which leads to an exponential increase of the parameters of models. Therein, the convolutional recurrent network as a type of widely used deep learning method is often employed to handle spatiotemporal data, e.g., traffic data. However, because of the large number of parameters in the model, the convolutional recurrent network needs to consume a lot of computing resources and time in the training process. To reduce the consumption of resources, we propose a sparse convolutional recurrent network with a sparse gating mechanism that is able to reduce the complexity of the network by an improved gate unit while keeping the performance of the model. We evaluate the performance of our proposed network on traffic flow datasets, and the experimental results show that the parameters of the model are significantly reduced under the condition of similar prediction accuracy compared with the traditional convolutional recurrent network.

Impact of Cirrhotic Cardiomyopathy Diagnosed According to Different Criteria on Patients with Cirrhosis Awaiting Liver Transplantation: A Retrospective Cohort Study.

Recently, the Cirrhotic Cardiomyopathy Consortium (Consortium) proposed criteria to replace the World Congress of Gastroenterology (WGO) criteria for cirrhotic cardiomyopathy (CCM) using contemporary echocardiography parameters. We assessed the impact of substituting WGO by Consortium criteria on the frequency of diagnosis and clinical outcomes in patients with cirrhosis awaiting liver transplantation (LT). Consecutive adults with cirrhosis approved for LT with echocardiography evaluation from January 2014 to December 2016 were screened. Patients with structural heart diseases were excluded. Two primary outcomes were: (1) frequency of CCM; (2) association of CCM with pre-transplant mortality. The secondary outcomes were pre-LT complications of acute kidney injury (AKI) and/or hepatic encephalopathy (HE), and post-LT mortality. Of 386 patients screened, 278 were included. 238 (85.6%) and 208 (74.8%) patients met Consortium and WGO criteria, respectively; 180 (64.7%) patients fulfilled both the criteria, while 12 (4.3%) patients had no evidence of CCM by either criterion. Pre-LT mortality rates in Consortium-CCM group were similar to the other groups (19.3% vs 20.2% vs 25.0%). The patients with advanced diastolic dysfunction (DD) per Consortium-CCM criteria had higher mortality than the other groups. The rates of pre-LT AKI/HE rates and post-LT mortality were similar in Consortium-CCM and WGO-CCM groups. The Consortium criteria do not impact the prevalence of CCM compared to WGO criteria and have similar predictive accuracy. Presence of advanced DD per the Consortium criteria increases the risk of pre-LT mortality and complications of AKI/HE. The patients with advanced DD could benefit from further monitoring and treatment.

Utility of Surgical Risk Calculators in Spine Surgery in Patients Aged Over 80 Years: Analysis of SpineSage and ACS NSQIP.

Retrospective cohort study. To validate the predictive accuracy of both the SpineSage and ACS-NSQIP surgical risk calculators in patients over the age of 80years, undergoing spine surgery for any reason. We included 210 consecutive patients treated with single-stage spine surgery at our institution between 2009 and 2019. The demographic details and preoperative characteristics of each patient were collected and reviewed for entry into both the SpineSage and ACS risk calculators. The estimated risk supplied by these calculators was compared to the observed rate of complications post-surgery. The main method of comparison was using receiver operating characteristic (ROC) curve analysis. Complications were identified in 51 patients (24%). Most patients underwent surgery for a degenerative cause (71%), with the majority of procedures performed on the lumbosacral spine (66%). Receiver operating characteristic (ROC) curves were calculated to compare the outcomes of each tool. Area under the curve (AUC) analysis showed similar predictive accuracy between SpineSage and ACS when predicting overall complications (0.688; P < .001 vs 0.634; P = .021). AUC analysis demonstrated that SpineSage had better predictive accuracy when estimating risk of major complications (0.778; P = .037 vs 0.675; P = .001). For the prediction of risks associated with spine surgery in those aged >80years, SpineSage appears to be preferable to the ACS-NSQIP surgical risk calculator in this single centre cohort, SpineSage was more accurate in predicting the risk of serious medical complications. The accuracy of both of these tools could still be improved upon.