Search

Cardiac magnetic resonance (CMR) image segmentation is an integral step in the analysis of cardiac function and diagnosis of heart related diseases. While recent deep learning-based approaches in automatic segmentation have shown great promise to alleviate the need for manual segmentation, most of these are not applicable to realistic clinical scenarios. This is largely due to training on mainly homogeneous datasets, without variation in acquisition, which typically occurs in multi-vendor and multi-site settings, as well as pathological data. Such approaches frequently exhibit a degradation in prediction performance, particularly on outlier cases commonly associated with difficult pathologies, artifacts and extensive changes in tissue shape and appearance. In this work, we present a model aimed at segmenting all three cardiac structures in a multi-center, multi-disease and multi-view scenario. We propose a pipeline, addressing different challenges with segmentation of such heterogeneous data, consisting of heart region detection, augmentation through image synthesis and a late-fusion segmentation approach. Extensive experiments and analysis demonstrate the ability of the proposed approach to tackle the presence of outlier cases during both training and testing, allowing for better adaptation to unseen and difficult examples. Overall, we show that the effective reduction of segmentation failures on outlier cases has a positive impact on not only the average segmentation performance, but also on the estimation of clinical parameters, leading to a better consistency in derived metrics.

A vast number of microarray datasets have been produced as a way to identify differentially expressed genes and gene expression signatures. A better understanding of these biological processes can help in the diagnosis and prognosis of diseases, as well as in the therapeutic response to drugs. However, most of the available datasets are composed of a reduced number of samples, leading to low statistical, predictive and generalization power. One way to overcome this problem is by merging several microarray datasets into a single dataset, which is typically a challenging task. Statistical methods or supervised machine learning algorithms are usually used to determine gene expression signatures. Nevertheless, statistical methods require an arbitrary threshold to be defined, and supervised machine learning methods can be ineffective when applied to high-dimensional datasets like microarrays. We propose a methodology to identify gene expression signatures by merging microarray datasets. This methodology uses statistical methods to obtain several sets of differentially expressed genes and uses supervised machine learning algorithms to select the gene expression signature. This methodology was validated using two distinct research applications: one using heart failure and the other using autism spectrum disorder microarray datasets. For the first, we obtained a gene expression signature composed of 117 genes, with a classification accuracy of approximately 98%. For the second use case, we obtained a gene expression signature composed of 79 genes, with a classification accuracy of approximately 82%. This methodology was implemented in R language and is available, under the MIT licence, at https://github.com/bioinformatics-ua/MicroGES.

The COVID-19 disease pandemic spread rapidly worldwide and caused extensive human death and financial losses. Therefore, finding accurate, accessible, and inexpensive methods for diagnosing the disease has challenged researchers. To automate the process of diagnosing COVID-19 disease through images, several strategies based on deep learning, such as transfer learning and ensemble learning, have been presented. However, these techniques cannot deal with noises and their propagation in different layers. In addition, many of the datasets already being used are imbalanced, and most techniques have used binary classification, COVID-19, from normal cases. To address these issues, we use the blind/referenceless image spatial quality evaluator to filter out inappropriate data in the dataset. In order to increase the volume and diversity of the data, we merge two datasets. This combination of two datasets allows multi-class classification between the three states of normal, COVID-19, and types of pneumonia, including bacterial and viral types. A weighted multi-class cross-entropy is used to reduce the effect of data imbalance. In addition, a fuzzy fine-tuned Xception model is applied to reduce the noise propagation in different layers. Quantitative analysis shows that our proposed model achieves 96.60% accuracy on the merged test set, which is more accurate than previously mentioned state-of-the-art methods.

Transgelin-2 (TG2) is a novel promising therapeutic target for the treatment of asthma as it plays an important role in relaxing airway smooth muscles and reducing pulmonary resistance in asthma. The compound TSG12 is the only reported TG2 agonist with in vivo anti-asthma activity. However, the dynamic behavior and ligand binding sites of TG2 and its binding mechanism with TSG12 remain unclear. In this study, we performed 12.6 μs molecular dynamics (MD) simulations for apo-TG2 and TG2-TSG12 complex, respectively. The results suggested that the apo-TG2 has 4 most populated conformations, and that its binding of the agonist could expand the conformation distribution space of the protein. The simulations revealed 3 potential binding sites in 3 most populated conformations, one of which is induced by the agonist binding. Free energy decomposition uncovered 8 important residues with contributions stronger than -1 kcal/mol. Computational alanine scanning for the important residues by 100 ns conventional MD simulation for each mutated TG2-TSG12 complexes demonstrated that E27, R49 and F52 are essential residues for the agonist binding. These results should be helpful to understand the dynamic behavior of TG2 and its binding mechanism with the agonist TSG12, which could provide some structural insights into the novel mechanism for anti-asthma drug development.

Human ether-a-go-go-related gene (hERG) channel blockade by small molecules is a big concern during drug development in the pharmaceutical industry. Failure or inhibition of hERG channel activity caused by drug molecules can lead to prolonging QT interval, which will result in serious cardiotoxicity. Thus, evaluating the hERG blocking activity of all these small molecular compounds is technically challenging, and the relevant procedures are expensive and time-consuming. In this study, we develop a novel deep learning predictive model named DMFGAM for predicting hERG blockers. In order to characterize the molecule more comprehensively, we first consider the fusion of multiple molecular fingerprint features to characterize its final molecular fingerprint features. Then, we use the multi-head attention mechanism to extract the molecular graph features. Both molecular fingerprint features and molecular graph features are fused as the final features of the compounds to make the feature expression of compounds more comprehensive. Finally, the molecules are classified into hERG blockers or hERG non-blockers through the fully connected neural network. We conduct 5-fold cross-validation experiment to evaluate the performance of DMFGAM, and verify the robustness of DMFGAM on external validation datasets. We believe DMFGAM can serve as a powerful tool to predict hERG channel blockers in the early stages of drug discovery and development.

The study of drug-target protein interaction is a key step in drug research. In recent years, machine learning techniques have become attractive for research, including drug research, due to their automated nature, predictive power, and expected efficiency. Protein representation is a key step in the study of drug-target protein interaction by machine learning, which plays a fundamental role in the ultimate accomplishment of accurate research. With the progress of machine learning, protein representation methods have gradually attracted attention and have consequently developed rapidly. Therefore, in this review, we systematically classify current protein representation methods, comprehensively review them, and discuss the latest advances of interest. According to the information extraction methods and information sources, these representation methods are generally divided into structure and sequence-based representation methods. Each primary class can be further divided into specific subcategories. As for the particular representation methods involve both traditional and the latest approaches. This review contains a comprehensive assessment of the various methods which researchers can use as a reference for their specific protein-related research requirements, including drug research.

In clinical applications, multi-dose scan protocols will cause the noise levels of computed tomography (CT) images to fluctuate widely. The popular low-dose CT (LDCT) denoising network outputs denoised images through an end-to-end mapping between an LDCT image and its corresponding ground truth. The limitation of this method is that the reduced noise level of the image may not meet the diagnostic needs of doctors. To establish a denoising model adapted to the multi-noise levels robustness, we proposed a novel and efficient modularized iterative network framework (MINF) to learn the feature of the original LDCT and the outputs of the previous modules, which can be reused in each following module. The proposed network can achieve the goal of gradual denoising, outputting clinical images with different denoising levels, and providing the reviewing physicians with increased confidence in their diagnosis. Moreover, a multi-scale convolutional neural network (MCNN) module is designed to extract as much feature information as possible during the network's training. Extensive experiments on public and private clinical datasets were carried out, and comparisons with several state-of-the-art methods show that the proposed method can achieve satisfactory results for noise suppression of LDCT images. In further comparisons with modularized adaptive processing neural network (MAP-NN), the proposed network shows superior step-by-step or gradual denoising performance. Considering the high quality of gradual denoising results, the proposed method can obtain satisfactory performance in terms of image contrast and detail protection as the level of denoising increases, which shows its potential to be suitable for a multi-dose levels denoising task.

The O6-methylguanine-DNA methyltransferase (MGMT) is a deoxyribonucleic acid (DNA) repairing enzyme that has been established as an essential clinical brain tumor biomarker for Glioblastoma Multiforme (GBM). Knowing the status of MGMT methylation biomarkers using multi-parametric MRI (mp-MRI) helps neuro-oncologists to analyze GBM and its treatment plan.The hand-crafted radiomics feature extraction of GBM's subregions, such as edema(ED), tumor core (TC), and enhancing tumor (ET) in the machine learning (ML) framework, was investigated using support vector machine(SVM), K-Nearest Neighbours (KNN), random forest (RF), LightGBM, and extreme gradient boosting (XGB). For tissue-level analysis of the promotor genes in GBM, we used the deep residual neural network (ResNet-18) with 3D architecture, followed by EfficientNet-based investigation for variants as B0 and B1. Lastly, we analyzed the fused deep learning (FDL) framework that combines ML and DL frameworks.Structural mp-MRI consisting of T1, T2, FLAIR, and T1GD having a size of 400 and 185 patients, respectively, for discovery and replication cohorts. Using the CV protocol in the ResNet-3D framework, MGMT methylation status prediction in mp-MRI gave the AUC of 0.753 (p 0.0001) and 0.72 (p 0.0001) for the discovery and replication cohort, respectively. We presented that the FDL is ∼7% superior to solo DL and ∼15% to solo ML.The proposed study aims to provide solutions for building an efficient predictive model of MGMT for GBM patients using deep radiomics features obtained from mp-MRI with the end-to-end ResNet-18 3D and FDL imaging signatures.

Understanding prognosis and mortality is critical for evaluating the treatment plan of patients. Advances in digital pathology and deep learning techniques have made it practical to perform survival analysis in whole slide images (WSIs). Current methods are usually based on a multi-stage framework which includes patch sampling, feature extraction and prediction. However, the random patch sampling strategy is highly unstable and prone to sampling non-ROI. Feature extraction typically relies on hand-crafted features or convolutional neural networks (CNNs) pre-trained on ImageNet, while the artificial error or domain gaps may affect the survival prediction performance. Besides, the limited information representation of local sampling patches will create a bottleneck limitation on the effectiveness of prediction. To address the above challenges, we propose a novel patch sampling strategy based on image information entropy and construct a Multi-Scale feature Fusion Network (MSFN) based on self-supervised feature extractor. Specifically, we adopt image information entropy as a criterion to select representative sampling patches, thereby avoiding the noise interference caused by random to blank regions. Meanwhile, we pretrain the feature extractor utilizing self-supervised learning mechanism to improve the efficiency of feature extraction. Furthermore, a global-local feature fusion prediction network based on the attention mechanism is constructed to improve the survival prediction effect of WSIs with comprehensive multi-scale information representation. The proposed method is validated by adequate experiments and achieves competitive results on both of the most popular WSIs survival analysis datasets, TCGA-GBM and TCGA-LUSC. Code and trained models are made available at: https://github.com/Mercuriiio/MSFN.

The main (Mpro) and papain-like (PLpro) proteases are highly conserved viral proteins essential for replication of the COVID-19 virus, SARS-COV-2. Therefore, a logical plan for producing new drugs against this pathogen is to discover inhibitors of these enzymes. Accordingly, the goal of the present work was to devise a computational approach to design, characterize, and select compounds predicted to be potent dual inhibitors - effective against both Mpro and PLpro. The first step employed LigDream, an artificial neural network, to create a virtual ligand library. Ligands with computed ADMET profiles indicating drug-like properties and low mammalian toxicity were selected for further study. Initial docking of these ligands into the active sites of Mpro and PLpro was done with GOLD, and the highest-scoring ligands were redocked with AutoDock Vina to determine binding free energies (ΔG). Compounds 89-00, 89-07, 89-32, and 89-38 exhibited favorable ΔG values for Mpro (-7.6 to -8.7 kcal/mol) and PLpro (-9.1 to -9.7 kcal/mol). Global docking of selected compounds with the Mpro dimer identified prospective allosteric inhibitors 89-00, 89-27, and 89-40 (ΔG -8.2 to -8.9 kcal/mol). Molecular dynamics simulations performed on Mpro and PLpro active site complexes with the four top-scoring ligands from Vina demonstrated that the most stable complexes were formed with compounds 89-32 and 89-38. Overall, the present computational strategy generated new compounds with predicted drug-like characteristics, low mammalian toxicity, and high inhibitory potencies against both target proteases to form stable complexes. Further preclinical studies will be required to validate the in silico findings before the lead compounds could be considered for clinical trials.

T cells are present in all stages of tumor formation and play an important role in the tumor microenvironment. We aimed to explore the expression profile of T cell marker genes, constructed a prognostic risk model based on these genes in Lung adenocarcinoma (LUAD), and investigated the link between this risk model and the immunotherapy response.We obtained the single-cell sequencing data of LUAD from the literature, and screened out 6 tissue biopsy samples, including 32,108 cells from patients with non-small cell lung cancer, to identify T cell marker genes in LUAD. Combined with TCGA database, a prognostic risk model based on T-cell marker gene was constructed, and the data from GEO database was used for verification. We also investigated the association between this risk model and immunotherapy response.Based on scRNA-seq data 1839 T-cell marker genes were identified, after which a risk model consisting of 9 gene signatures for prognosis was constructed in combination with the TCGA dataset. This risk model divided patients into high-risk and low-risk groups based on overall survival. The multivariate analysis demonstrated that the risk model was an independent prognostic factor. Analysis of immune profiles showed that high-risk groups presented discriminative immune-cell infiltrations and immune-suppressive states. Risk scores of the model were closely correlated with Linoleic acid metabolism, intestinal immune network for IgA production and drug metabolism cytochrome P450.Our study proposed a novel prognostic risk model based on T cell marker genes for LUAD patients. The survival of LUAD patients as well as treatment outcomes may be accurately predicted by the prognostic risk model, and make the high-risk population present different immune cell infiltration and immunosuppression state.

Due to its high transmissibility, Omicron BA.1 ousted the Delta variant to become a dominating variant in late 2021 and was replaced by more transmissible Omicron BA.2 in March 2022. An important question is which new variants will dominate in the future. Topology-based deep learning models have had tremendous success in forecasting emerging variants in the past. However, topology is insensitive to homotopic shape evolution in virus-human protein-protein binding, which is crucial to viral evolution and transmission. This challenge is tackled with persistent Laplacian, which is able to capture both the topological change and homotopic shape evolution of data. Persistent Laplacian-based deep learning models are developed to systematically evaluate variant infectivity. Our comparative analysis of Alpha, Beta, Gamma, Delta, Lambda, Mu, and Omicron BA.1, BA.1.1, BA.2, BA.2.11, BA.2.12.1, BA.3, BA.4, and BA.5 unveils that Omicron BA.2.11, BA.2.12.1, BA.3, BA.4, and BA.5 are more contagious than BA.2. In particular, BA.4 and BA.5 are about 36% more infectious than BA.2 and are projected to become new dominant variants by natural selection. Moreover, the proposed models outperform the state-of-the-art methods on three major benchmark datasets for mutation-induced protein-protein binding free energy changes. Our key projection about BA4 and BA.5's dominance made on May 1, 2022 (see arXiv:2205.00532) became a reality in late June 2022.

SHP2 (Src homology-2 domain-containing protein tyrosine phosphatase-2) is a cytoplasmic protein -tyrosine phosphatase encoded by the gene PTPN11. It plays a crucial role in regulating cell growth and differentiation. Specifically, SHP2 is an oncoprotein associated with developmental pathologies and several different cancer types, including gastric, leukemia and breast cancer and is of great therapeutic interest. Given these roles, current research efforts have focused on developing SHP2 inhibitors. Allosteric SHP2 inhibitors have been shown to be more selective and pharmacologically appealing compared to competitive catalytic inhibitors targeting SHP2. Nevertheless, there remains a need for novel allosteric inhibitor scaffolds targeting SHP2 to develop compounds with improved selectivity, cell permeability, and bioavailability. Towards this goal, this study applied various computational tools to screen over 6 million compounds against the allosteric site within SHP2. The top-ranked hits from our in-silico screening were validated using protein thermal shift and biolayer interferometry assays, revealing three potent compounds. Kinetic binding assays were employed to measure the binding affinities of the top-ranked compounds and demonstrated that they all bind to SHP2 with a nanomolar affinity. Hence the compounds and the computational workflow described herein provide an effective approach for identifying and designing a generation of improved allosteric inhibitors of SHP2.

Feature selection (FS) is a popular data pre-processing technique in machine learning to extract the optimal features to maintain or increase the classification accuracy of the dataset, which is a combinatorial optimization problem, requiring a powerful optimizer to obtain the optimum subset. The equilibrium optimizer (EO) is a recent physical-based metaheuristic algorithm with good performance for various optimization problems, but it may encounter premature or the local convergence in feature selection. This work presents a self-adaptive quantum EO with artificial bee colony for feature selection, named SQEOABC. In the proposed algorithm, the quantum theory and the self-adaptive mechanism are employed into the updating rule of EO to enhance convergence, and the updating mechanism from the artificial bee colony is also incorporated into EO to achieve appropriate FS solutions. In the experiments, 25 benchmark datasets from the UCI repository are investigated to verify SQEOABC, which is compared with several state-of-the-art metaheuristic algorithms and the variants of EO. The statistical results of fitness values and accuracy demonstrate that SQEOABC has better performance than the compared algorithms and the variants of EO. Finally, a real-world FS problem from COVID-19 illustrates the effectiveness and superiority of SQEOABC.

Rapid advances in single-cell transcriptome analysis provide deeper insights into the study of tissue heterogeneity at the cellular level. Unsupervised clustering can identify potential cell populations in single-cell RNA-sequencing (scRNA-seq) data, but fail to further determine the identity of each cell. Existing automatic annotation methods using scRNA-seq data based on machine learning mainly use single feature set and single classifier. In view of this, we propose a Weighted Ensemble classification framework for Cell Type Annotation, named scWECTA, which improves the accuracy of cell type identification. scWECTA uses five informative gene sets and integrates five classifiers based on soft weighted ensemble framework. And the ensemble weights are inferred through the constrained non-negative least squares. Validated on multiple pairs of scRNA-seq datasets, scWECTA is able to accurately annotate scRNA-seq data across platforms and across tissues, especially for imbalanced data containing rare cell types. Moreover, scWECTA outperforms other comparable methods in balancing the prediction accuracy of common cell types and the unassigned rate of non-common cell types at the same time. The source code of scWECTA is freely available at https://github.com/ttren-sc/scWECTA.

One of the major challenges in drug development is having acceptable levels of efficacy and safety throughout all the phases of clinical trials followed by the successful launch in the market. While there are many factors such as molecular properties, toxicity parameters, mechanism of action at the target site, etc. that regulates the therapeutic action of a compound, a holistic approach directed towards data-driven studies will invariably strengthen the predictive toxicological sciences. Our quest for the current study is to find out various reasons as to why an investigational candidate would fail in the clinical trials after multiple iterations of refinement and optimization. We have compiled a dataset that comprises of approved and withdrawn drugs as well as toxic compounds and essentially have used time-split based approach to generate the training and validation set. Five highly robust and scalable machine learning binary classifiers were used to develop the predictive models that were trained with features like molecular descriptors and fingerprints and then validated rigorously to achieve acceptable performance in terms of a set of performance metrics. The mean AUC scores for all the five classifiers with the hold-out test set were obtained in the range of 0.66-0.71. The models were further used to predict the probability score for the clinical candidate dataset. The top compounds predicted to be toxic were analyzed to estimate different dimensions of toxicity. Apparently, through this study, we propose that with the appropriate use of feature extraction and machine learning methods, one can estimate the likelihood of success or failure of investigational drugs candidates thereby opening an avenue for future trends in computational toxicological studies. The models developed in the study can be accessed at https://github.com/gnsastry/predicting_clinical_trials.git.

Liver tumours are diseases with high morbidity and high deterioration probabilities, and accurate liver area segmentation from computed tomography (CT) scans is a prerequisite for quick tumour diagnosis. While 2D network segmentation methods can perform segmentation with lower device performance requirements, they often discard the rich 3D spatial information contained in CT scans, limiting their segmentation accuracy. Hence, a deep residual attention-based U-shaped network (DRAUNet) with a biplane joint method for liver segmentation is proposed in this paper, where the biplane joint method introduces coronal CT slices to assist the transverse slices with segmentation, incorporating more 3D spatial information into the segmentation results to improve the segmentation performance of the network. Additionally, a novel deep residual block (DR block) and dual-effect attention module (DAM) are introduced in DRAUNet, where the DR block has deeper layers and two shortcut paths. The DAM efficiently combines the correlations of feature channels and the spatial locations of feature maps. The DRAUNet with the biplane joint method is tested on three datasets, Liver Tumour Segmentation (LiTS), 3D Image Reconstruction for Comparison of Algorithms Database (3DIRCADb), and Segmentation of the Liver Competition 2007 (Sliver07), and it achieves 97.3%, 97.4%, and 96.9% Dice similarity coefficients (DSCs) for liver segmentation, respectively, outperforming most state-of-the-art networks; this strongly demonstrates the segmentation performance of DRAUNet and the ability of the biplane joint method to obtain 3D spatial information from 3D images.

In recent years, with the gradual increase in pancancer-related research, more attention has been given to the field of pancancer metastasis. However, the molecular mechanism of pancancer metastasis is very unclear, and identification methods for pancancer metastasis-related genes are still lacking. In view of this research status, we developed a novel pipeline to identify pancancer metastasis-related genes based on compound constrained nonnegative matrix factorization (CCNMF). To solve the above problems, the following modules were designed. A correntropy operator and feature similarity fusion (FSF) were first adopted to process the multiomics features of genes; thus, the influences caused by irrelevant biomolecular patterns, manifested as non-Gaussian noise, were minimized. CCNMF was then adopted to handle the above features with compound constraints consisting of a gene relation network and a "metastasis-related" gene set, which maximizes the biological interpretability of the metafeatures generated by NMF. Since a negative set of pancancer "metastasis-related" genes could hardly be obtained, semisupervised analyses were performed on gene features acquired by each step in our pipeline to examine our method's effect. 83% of the 236 candidates identified by the above method were associated with the metastasis of one or more cancers, 71.9% candidates were identified immune-related in pancancer in addition to the hallmark genes. Our study provides an effective and interpretable method for identifying metastasis-related as well as immune-related genes, and the method is successfully applied to TCGA pancancer data.

The human immunodeficiency virus (HIV) connects to the cluster of differentiation (CD4) and any of the entry co-receptors (CCR5 and CXCR4); followed by unloading the viral genome, reverse transcriptase, and integrase enzymes within the host cell. The co-receptors facilitate the entry of virus and vital enzymes, leading to replication and pre-maturation of viral particles within the host. The protease enzyme transforms the immature viral vesicles into the mature virion. The pivotal role of co-receptors and enzymes in homeostasis and growth makes the crucial target for anti-HIV drug discovery, and the availability of X-ray crystal structures is an asset. Here, we used the machine intelligence-driven framework (A-HIOT) to identify and optimize target-based potential hit molecules for five significant protein targets from the ZINC15 database (natural products dataset). Following validation with dynamic motion behavior analysis and molecular dynamics simulation, the optimized hits were evaluated using in silico ADMET filtration. Furthermore, three molecules were screened, optimized, and validated: ZINC00005328058 for CCR5 and protease, ZINC000254014855 for CXCR4 and integrase, and ZINC000000538471 for reverse transcriptase. In clinical trials, the ZINC000254014855 and ZINC000254014855 were passed in primary screens for vif-HIV-1, and we reported the specific receptor as well as interactions. As a result, the validated molecules may be investigated further in experimental studies targeting specific receptors in order to design and synergize an anti-HIV regimen.

Despite the numerous studies on extubation readiness assessment for patients who are invasively ventilated in the intensive care unit, a 10-15% extubation failure rate persists. Although breathing variability has been proposed as a potential predictor of extubation failure, it is mainly assessed using simple statistical metrics applied to basic respiratory parameters. Therefore, the complex pattern of breathing variability conveyed by continuous ventilation waveforms may be underexplored.Here, we aimed to develop novel breathing variability indices to predict extubation failure among invasively ventilated patients. First, breath-to-breath basic and comprehensive respiratory parameters were computed from continuous ventilation waveforms 1 h before extubation. Subsequently, the basic and advanced variability methods were applied to the respiratory parameter sequences to derive comprehensive breathing variability indices, and their role in predicting extubation failure was assessed. Finally, after reducing the feature dimensionality using the forward search method, the combined effect of the indices was evaluated by inputting them into the machine learning models, including logistic regression, random forest, support vector machine, and eXtreme Gradient Boosting (XGBoost).The coefficient of variation of the dynamic mechanical power per breath (CV-MPThese results suggest that the proposed novel breathing variability indices can improve extubation failure prediction in invasively ventilated patients.

In this paper, we proposed an enhanced reptile search algorithm (RSA) for global optimization and selected optimal thresholding values for multilevel image segmentation. RSA is a recent metaheuristic optimization algorithm depending on the hunting behavior of crocodiles. RSA is inclined to inadequate diversity, local optima, and unbalanced exploitation abilities as other metaheuristic algorithms. The RUNge Kutta optimizer (RUN) is a novel metaheuristic algorithm that has demonstrated effectiveness in solving real-world optimization problems. The enhanced solution quality (ESQ) in RUN utilizes the thus-far best solution to promote the quality of solutions, improve the convergence speed, and effectively balance the exploration and exploitation steps. Also, the Scale factor (SF) has a randomized adaptation nature, which helps RUN in further improving the exploration and exploitation steps. This parameter ensures a smooth transition from exploration to exploitation. In order to mitigate the drawbacks of the RSA algorithm, this paper proposed a modified RSA (mRSA), which combines the RSA algorithm with the RUN. The ESQ mechanism and the scale factor boost the original RSA's performance, enhance convergence speed, bypass local optimum, and enhance the balance between exploitation and exploration. The validity of mRSA was verified using two experimental sequences. First, we applied mRSA to CEC'2020 benchmark functions of various types and dimensions, showing that mRSA has more robust search capabilities than the original RSA and popular counterpart algorithms concerning statistical, convergence, and diversity measurements. The second experiment evaluated mRSA for a real-world application to solve magnetic resonance imaging (MRI) brain image segmentation. Overall experimental results confirm that the mRSA has a strong optimization ability. Also, mRSA method is a more successful multilevel thresholding segmentation and outperforms comparison methods according to different performance measures.

The automatic extraction of procedural surgical knowledge from surgery manuals, academic papers or other high-quality textual resources, is of the utmost importance to develop knowledge-based clinical decision support systems, to automatically execute some procedure's step or to summarize the procedural information, spread throughout the texts, in a structured form usable as a study resource by medical students. In this work, we propose a first benchmark on extracting detailed surgical actions from available intervention procedure textbooks and papers. We frame the problem as a Semantic Role Labeling task. Exploiting a manually annotated dataset, we apply different Transformer-based information extraction methods. Starting from RoBERTa and BioMedRoBERTa pre-trained language models, we first investigate a zero-shot scenario and compare the obtained results with a full fine-tuning setting. We then introduce a new ad-hoc surgical language model, named SurgicBERTa, pre-trained on a large collection of surgical materials, and we compare it with the previous ones. In the assessment, we explore different dataset splits (one in-domain and two out-of-domain) and we investigate also the effectiveness of the approach in a few-shot learning scenario. Performance is evaluated on three correlated sub-tasks: predicate disambiguation, semantic argument disambiguation and predicate-argument disambiguation. Results show that the fine-tuning of a pre-trained domain-specific language model achieves the highest performance on all splits and on all sub-tasks. All models are publicly released.

Breast cancer is the main cancer type with more than 2.2 million cases in 2020, and is the principal cause of death in women; with 685000 deaths in 2020 worldwide. The estrogen receptor is involved at least in 70% of breast cancer diagnoses, and the agonist and antagonist properties of the drug in this receptor play a pivotal role in the control of this illness. This work evaluated the agonist and antagonist mechanisms of 30 cannabinoids by employing molecular docking and dynamic simulations. Compounds with docking scores -8 kcal/mol were analyzed by molecular dynamic simulation at 300 ns, and relevant insights are given about the protein's structural changes, centered on the helicity in alpha-helices H3, H8, H11, and H12. Cannabicitran was the cannabinoid that presented the best relative binding-free energy (-34.96 kcal/mol), and based on rational modification, we found a new natural-based compound with relative binding-free energy (-44.83 kcal/mol) better than the controls hydroxytamoxifen and acolbifen. Structure modifications that could increase biological activity are suggested.

Over the last decades, molecular signatures have attracted extensive attention in cancer research. However, most of the reported biomarkers show a weak distinguishing ability in predicting the survival risks of patients. Actually, univariate analysis is generally considered in regression analysis, which makes the existing statistical methods ineffective. Furthermore, there is too much human involvement in the ways of classifying patients with high and low risk. Last but not least, the participation of therapy after conservative surgery also makes the survival analysis more complex. In order to solve these problems, we propose a solid method of feature selection which combines top-down and bottom-up strategies. The top-down strategy is to randomly extract some genes each time and select candidate genes through cumulative voting. The bottom-up strategy is to fully enumerate the selected genes and to use a clustering algorithm to classify samples. We analyzed glioblastoma data from the Cancer Genome Atlas (TCGA) and got candidate signatures. The results of simulation data, as well as an independent test set the Chinese Glioma Genome Atlas (CGGA), verified the reliability of the method and validity of the selected features.

The number of SARS-CoV-2 spike Receptor Binding Domain (RBD) with multiple amino acid mutations is huge due to random mutations and combinatorial explosions, making it almost impossible to experimentally determine their binding affinities to human angiotensin-converting enzyme 2 (hACE2). Although computational prediction is an alternative way, there is still no online platform to predict the mutation effect of RBD on the hACE2 binding affinity until now. In this study, we developed a free online platform based on deep learning models, namely D3AI-Spike, for quickly predicting binding affinity between spike RBD mutants and hACE2. The models based on CNN and CNN-RNN methods have the concordance index of around 0.8. Overall, the test results of the models are in agreement with the experimental data. To further evaluate the prediction power of D3AI-Spike, we predicted and experimentally determined the binding affinity of a VUM (variants under monitoring) variant IHU (B.1.640.2), which has fourteen amino acid substitutions, including N501Y and E484K, and 9 deletions located in the spike protein. The predicted average affinity score for wild-type RBD and IHU to hACE2 are 0.483 and 0.438, while the determined K

Increasing interest has been attracted in deciphering the potential disease pathogenesis through lncRNA-disease association (LDA) prediction, regarding to the diverse functional roles of lncRNAs in genome regulation. Whilst, computational models and algorithms benefit systematic biology research, even facilitate the classical biological experimental procedures. In this review, we introduce representative diseases associated with lncRNAs, such as cancers, cardiovascular diseases, and neurological diseases. Current publicly available resources related to lncRNAs and diseases have also been included. Furthermore, all of the 64 computational methods for LDA prediction have been divided into 5 groups, including machine learning-based methods, network propagation-based methods, matrix factorization- and completion-based methods, deep learning-based methods, and graph neural network-based methods. The common evaluation methods and metrics in LDA prediction have also been discussed. Finally, the challenges and future trends in LDA prediction have been discussed. Recent advances in LDA prediction approaches have been summarized in the GitHub repository at https://github.com/sheng-n/lncRNA-disease-methods.

A brain tumor is a dynamic system in which cells develop rapidly and abnormally, as is the case with most cancers. Cancer develops in the brain or inside the skull when aberrant and odd cells proliferate in the brain. By depriving the healthy cells of leisure, nutrition, and oxygen, these aberrant cells eventually cause the healthy cells to perish. This article investigated the development of glioma cells in treating brain tumors. Mathematically, reaction-diffusion models have been developed for brain glioma growth to quantify the diffusion and proliferation of the tumor cells within brain tissues. This study presents the formulation the two-stage successive over-relaxation (TSSOR) algorithm based on the finite difference approximation for solving the treated brain glioma model to predict glioma cells in treating the brain tumor. Also, the performance of TSSOR method is compared to the Gauss-Seidel (GS) and two-stage Gauss-Seidel (TSGS) methods in terms of the number of iterations, the amount of time it takes to process the data, and the rate at which glioma cells grow the fastest. The implementation of the TSSOR, TSGS, and GS methods predicts the growth of tumor cells under the treatment protocol. The results show that the number of glioma cells decreased initially and then increased gradually by the next day. The computational complexity analysis is also used and concludes that the TSSOR method is faster compared to the TSGS and GS methods. According to the results of the treated glioma development model, the TSSOR approach reduced the number of iterations by between 8.0 and 71.95%. In terms of computational time, the TSSOR approach is around 1.18-76.34% faster than the TSGS and GS methods.

A major challenge in artificial intelligence based ECG diagnosis lies that it is difficult to obtain sufficient annotated training samples for each rhythm type, especially for rare diseases, which makes many approaches fail to achieve the desired performance with limited ECG records. In this paper, we propose a Meta Siamese Network (MSN) based on metric learning to achieve high accuracy for automatic ECG arrhythmias diagnosis with limited ECG records. First, the ECG signals from three different ECG datasets are preprocessed through resampling, wavelet denoising, R-wave localization, heartbeat segmentation and Z-score normalization. Then, an ECG dataset with limited records is constructed to verify the performance of the proposed model and explore variation of model performance with the sample size. Second, a metric-based meta-learning framework is proposed to address the challenge of few-shot learning for automatic ECG diagnosis of cardiac arrhythmia, and siamese network is employed to achieve arrhythmia diagnosis based on similarity metric. Finally, the N-way K-shot meta-testing strategy is proposed based on the siamese network with double inputs, and the experimental results demonstrate that the proposed strategy can effectively improve the robustness of the proposed model.

Pre-processing is widely applied in medical image analysis to remove the interference information. However, the existing pre-processing solutions mainly encounter two problems: (i) it is heavily relied on the assistance of clinical experts, making it hard for intelligent CAD systems to deploy quickly; (ii) due to the personnel and information barriers, it is difficult for medical institutions to conduct the same pre-processing operations, making a deep model that performs well on a specific medical institution difficult to achieve similar performances on the same task in other medical institutions. To overcome these problems, we propose a deep-reinforcement-learning-based task-oriented homogenized automatic pre-processing (DRL-HAPre) framework to overcome these two problems. This framework utilizes deep reinforcement learning techniques to learn a policy network to automatically and adaptively select the optimal pre-processing operations for the input medical images according to different analysis tasks, thus helping the intelligent CAD system to achieve a rapid deployment (i.e., painless) and maintain a satisfactory performance (i.e., accurate) among different medical institutes. To verify the effectiveness and advantages of the proposed DRL-HAPre framework, we further develop a homogenized automatic pre-processing model based on the DRL-HAPre framework to realize the automatic pre-processing of key region selection (called HAPre-KRS) in the pneumonia image classification task. Extensive experimental studies are conducted on three pediatric pneumonia classification datasets with different image qualities, and the results show that: (i) There does exist a hard-to-reproduce problem in clinical practices and the fact that having different medical image qualities in different medical institutes is an important reason for the existing of hard-to-reproduce problem, so it is compelling to propose homogenized automatic pre-processing method. (ii) The proposed HAPre-KRS model and DRL-HAPre framework greatly outperform three kinds of state-of-the-art baselines (i.e., pre-processing, attention and pneumonia baseline), and the lower the medical image quality, the greater the improvements of using our HAPre-KRS model and DRL-HAPre framework. (iii) With the help of homogenized pre-processing, HAPre-KRS (and DRL-HAPre framework) can greatly avoid performance degradation in real-world cross-source applications (i.e., thus overcoming the hard-to-reproduce problem).

More than 150 genes are involved in amyotrophic lateral sclerosis (ALS), with superoxide dismutase 1 (SOD1) being one of the most studied. Mutations in SOD1 gene, which encodes the enzyme SOD1 is the second most prevalent and studied cause of familial ALS. SOD1 is a ubiquitous, homodimeric metalloenzyme that forms a critical component of the cellular defense against reactive oxygen species. Several mutations in the SOD1 enzyme cause misfolding, dimerization instability, and increased aggregate formation in ALS. However, there is a lack of information on the dimerization of SOD1 monomers and the mechanistic underpinnings on how the pathogenic mutations disrupt the dimerization mechanism. Here, we presented microsecond-scale molecular dynamics (MD) simulations to unravel how interface-based mutations compromise SOD1 dimerization and provide mechanistic understanding into the corresponding process using WT and three interface-based mutant systems (A4V, T54R, and I113T). Structural stability analysis showed that the mutant systems displayed disparate variations in the catalytic sites which may directly alter the stability and activity of the SOD1 enzyme. Based on the dynamic network analysis and principal component analysis, it has been identified that the mutations weakened the correlated motions along the dimer interface and altered the protein conformational behavior, thus weakening the stability of dimer formation. Moreover, the simulation results identified crucial residues such as G51, D52, G114, I151, and Q153 in establishing the dimerization interaction network, which were weakened or absent in the presence of interfacial mutants. Surface potential analysis on mutant systems also displayed changes in the dimerization potential, thus showing the unfavorable dimer formation. Furthermore, network analysis identified the hotspot residues necessary for SOD1 signal transduction which were surprisingly found in the catalytic sites rather than the anticipated dimerization interface.

Alzheimer's disease is a significant public health issue. Recent studies have shown that the gut microbiota plays a vital role in the onset and development of Alzheimer's disease. However, the potential role of the gut microbiota and the associated metabolic characteristics require further elucidation.The gut microbial compositions of APP/PS1 mice were analyzed using 16S rRNA gene sequencing. Metabolomics was used to characterize changes in metabolic profiles in feces, serum, and cortex. A multi-omics approach investigated the potential associations between gut microbes and metabolites.The gut microbiota composition was markedly different between APP/PS1 mice and normal mice. Metabolomic analysis identified 253 fecal metabolites, 16 serum metabolites, and 123 cortical metabolites that were differentially abundant in APP/PS1 that may be potential biomarkers of AD. Nearly half of these metabolites were lipids. A combined analysis of the three sample types showed a correlation between fecal fatty acids and glycerolipids, serum glycerophospholipids, and cortical fatty acids. Furthermore, our study showed that Marinifilaceae and Akkermansiaceae were closely related to these lipids and lipid-like molecules, particularly fatty acids and glycerophospholipids.Our study highlighted the interactions between the gut microbiome and the fecal, serum, and cortical metabolomes. This interaction provides a new direction for further exploring the link between gut microbiota composition and metabolism in Alzheimer's disease.

Recent outbreak of monkeypox disease commenced in April 2022, and on May 7, the first confirmed case was reported. The world health organization then designated monkeypox disease as a public health emergency of international outrage on July 23, after it spread to 70 non-endemic nations in less than 15 days. This catastrophic viral infection encourages the development of antiviral therapeutics due to the lack of specific treatments with negligible adverse effects. This analysis developed a highly immunogenic multiepitope subunit vaccine against the monkeypox virus using an in silico translational vaccinomics technique. Highly antigenic B cell and T cell (HTL and CTL) epitopes were predicted and conjugated with the help of unique linkers. An adjuvant (β-defensin) and a pan-HLA DR sequence were attached at the vaccine construct's N-terminal to invoke a robust immunological response. Additionally, physiochemical, allergic, toxic, and antigenic properties were anticipated. Interactions between the vaccine candidate and the TLR3 demonstrated that the vaccine candidate triggers a robust immunological response. Finally, the stability is confirmed by the molecular dynamics study. In contrast, the modified vaccine candidate's ability to produce a protective immune response were verified by an immune dynamics simulation study conducted via C-ImmSim server. This study validates the generation of B cell, Th cell, and Tc cell populations as well as the production of IFN-γ.

Mortality prediction is an important task in intensive care unit (ICU) for quantifying the severity of patients' physiological condition. Currently, scoring systems are widely applied for mortality prediction, while the performance is unsatisfactory in many clinical conditions due to the non-specificity and linearity characteristics of the used model. As the availability of the large volume of data recorded in electronic health records (EHRs), deep learning models have achieved state-of-art predictive performance. However, deep learning models are hard to meet the requirement of explainability in clinical conditions. Hence, an explainable Knowledge Distillation method with XGBoost (XGB-KD) is proposed to improve the predictive performance of XGBoost while supporting better explainability.In this method, we first use outperformed deep learning teacher models to learn the complex patterns hidden in high-dimensional multivariate time series data. Then, we distill knowledge from soft labels generated by the ensemble of teacher models to guide the training of XGBoost student model, whose inputs are meaningful features obtained from feature engineering. Finally, we conduct model calibration to obtain predicted probabilities reflecting the true posterior probabilities and use SHapley Additive exPlanations (SHAP) to obtain insights about the trained model.We conduct comprehensive experiments on MIMIC-III dataset to evaluate our method. The results demonstrate that our method achieves better predictive performance than vanilla XGBoost, deep learning models and several state-of-art baselines from related works. Our method can also provide intuitive explanations.Our method is useful for improving the predictive performance of XGBoost by distilling knowledge from deep learning models and can provide meaningful explanations for predictions.

Uterine carcinosarcoma (UCS) is an invasive variant of endometrial cancer. The complicated heterogeneity and low frequency of UCS suggest the relevant research is lack. There is an urgent need to further explore the pathogenic mechanism and identify new biomarkers of UCS from different angels to improve its diagnosis and prognosis.This study is to explore the importance of alternative splicing (AS) events in UCS, construct AS-based prognosis model and excavate key splicing factors (SFs).UCS related gene transcriptome data and AS events data were collected from The Cancer Genome Atlas (TCGA) and TCGA SpliceSeq database. The AS events related to survival were determined by Cox regression analysis, Least absolute shrinkage and selection operator (Lasso) regression analysis and optimal subset analysis. The corresponding risk score was calculated and its efficiency on prognosis was evaluated by Kaplan-Meier (K-M) survival estimate and validated by the receiver operating characteristic (ROC) curve. The prognosis model was constructed with risk score and clinic characters as independent variables to predict patients' survival. On the other hand, Kendall test was applied to inspect the correlation between the SFs and the prognosis-related AS events and a AS-SF network was constructed. Finally, the key SFs were screened through network nodes analysis and survival analysis.Seven AS events the most related to survival were detected and the risk score was obtained. K-M survival estimate and ROC curve validation suggested the risk score was effective. Then Cox model was constructed based on the risk score and a nomogram model was obtained which provided the highest prediction accuracy of 95%. Through the AS-SF network analysis, 16 SFs were screened, among which four survival-related SFs were eventually obtained.The prognosis model could predict the survival rate of UCS patients by their clinical characters and AS-based risk score. And four newly discovered SFs could reveal the molecular mechanism of UCS and act as the potential drug targets and prognosis biomarkers.

Fatigue driving is one of the leading causes of traffic accidents, so fatigue driving detection technology plays a crucial role in road safety. The physiological information-based fatigue detection methods have the advantage of objectivity and accuracy. Among many physiological signals, EEG signals are considered to be the most direct and promising ones. Most traditional methods are challenging to train and do not meet real-time requirements. To this end, we propose an end-to-end temporal and graph convolution-based (MATCN-GT) fatigue driving detection algorithm. The MATCN-GT model consists of a multi-scale attentional temporal convolutional neural network block (MATCN block) and a graph convolutional-Transformer block (GT block). Among them, the MATCN block extracts features directly from the original EEG signal without a priori information, and the GT block processes the features of EEG signals between different electrodes. In addition, we design a multi-scale attention module to ensure that valuable information on electrode correlations will not be lost. We add a Transformer module to the graph convolutional network, which can better capture the dependencies between long-distance electrodes. We conduct experiments on the public dataset SEED-VIG, and the accuracy of the MATCN-GT model reached 93.67%, outperforming existing algorithms. Furthermore, compared with the traditional graph convolutional neural network, the GT block has improved the accuracy rate by 3.25%. The accuracy of the MATCN block on different subjects is higher than the existing feature extraction methods.

Liver Ultrasound (US) or sonography is popularly used because of its real-time output, low-cost, ease-of-use, portability, and non-invasive nature. Segmentation of real-time liver US is essential for diagnosing and analyzing liver conditions (e.g., hepatocellular carcinoma (HCC)), assisting the surgeons/radiologists in therapeutic procedures. In this paper, we propose a method using a modified Pyramid Scene Parsing (PSP) module in tuned neural network backbones to achieve real-time segmentation without compromising the segmentation accuracy. Considering widespread noise in US data and its impact on outcomes, we study the impact of pre-processing and the influence of loss functions on segmentation performance. We have tested our method after annotating a publicly available US dataset containing 2400 images of 8 healthy volunteers (link to the annotated dataset is provided); the results show that the Dense-PSP-UNet model achieves a high Dice coefficient of 0.913±0.024 while delivering a real-time performance of 37 frames per second (FPS).

Cancer patients show heterogeneous phenotypes and very different outcomes and responses even to common treatments, such as standard chemotherapy. This state-of-affairs has motivated the need for the comprehensive characterization of cancer phenotypes and fueled the generation of large omics datasets, comprising multiple omics data reported for the same patients, which might now allow us to start deciphering cancer heterogeneity and implement personalized therapeutic strategies. In this work, we performed the analysis of four cancer types obtained from the latest efforts by The Cancer Genome Atlas, for which seven distinct omics data were available for each patient, in addition to curated clinical outcomes. We performed a uniform pipeline for raw data preprocessing and adopted the Cancer Integration via MultIkernel LeaRning (CIMLR) integrative clustering method to extract cancer subtypes. We then systematically review the discovered clusters for the considered cancer types, highlighting novel associations between the different omics and prognosis.