Independent Datasets Research Articles

RDR100: A Robust Computational Method for Identification of Krüppel-like Factors

Background: Krüppel-like factors (KLFs) are a family of transcription factors containing zinc fingers that regulate various cellular processes. KLF proteins are associated with human diseases, such as cancer, cardiovascular diseases, and metabolic disorders. The KLF family consists of 18 members with diverse expression profiles across numerous tissues. Accurate identification and annotation of KLF proteins is crucial, given their involvement in important biological functions. Although experimental approaches can identify KLF proteins precisely, large-scale identification is complicated, slow, and expensive. Methods: In this study, we developed RDR100, a novel random forest (RF)-based framework for predicting KLF proteins based on their primary sequences. First, we identified the optimal encodings for ten different features using a recursive feature elimination approach, and then trained their respective model using five distinct machine learning (ML) classifiers. Results: The performance of all models was assessed using independent datasets, and RDR100 was selected as the final model based on its consistent performance in cross-validation and independent evaluation. Conclusion: Our results demonstrate that RDR100 is a robust predictor of KLF proteins. RDR100 web server is available at https://procarb.org/RDR100/.

Metabolic pathway-based subtyping reveals distinct microenvironmental states associated with diffuse large B-cell lymphoma outcomes.

Diffuse large B-cell lymphoma (DLBCL) is a biologically and clinically heterogeneous disease that requires personalized clinical treatment. Assigning patients to different risk categories and cytogenetic abnormality and genetic mutation groups has been widely applied for prognostic stratification of DLBCL. Increasing evidence has demonstrated that dysregulated metabolic processes contribute to the initiation and progression of DLBCL. Metabolic competition within the tumor microenvironment is also known to influence immune cell metabolism. However, metabolism- and immune-related stratification has not been established. Here, 1660 genes involved in 84 metabolic pathways were selected and tested to establish metabolic clusters (MECs) of DLBCL. MECs established based on independent lymphoma datasets distinguished different survival outcomes. The CIBERSORT algorithm and EcoTyper were applied to quantify the relative abundance of immune cell types and identify variation in cell states for 13 lineages comprising the tumor micro environment among different MECs, respectively. Functional characterization showed that MECs were an indicator of the immune microenvironment and correlated with distinctive mutational characteristics and oncogenic signaling pathways. The novel immune-related MECs exhibited promising clinical prognostic value and potential for informing DLBCL treatment decisions.

Compulsivity-related behavioral features of problematic usage of the internet: A scoping review of paradigms, progress, and perspectives.

Compulsivity contributes to the development and maintenance of multiple addictive disorders. However, the relationship between compulsivity-related cognitive features and problematic usage of the internet (PUI), an umbrella term for various internet use disorders/interfering behaviors, remains largely unclear, partly due to the multidimensional nature of compulsivity. This scoping review utilized a four-domain framework of compulsivity to consider this topic and aimed to summarize available evidence on compulsivity-related neuropsychological characteristics in PUI based on this framework. A systematic literature search was conducted by applying the combination of search term to the search engines of PubMed, PsycINFO and Web of Science. A four-domain framework of compulsivity, involving cognitive flexibility, set-shifting, attentional bias, and habit learning, was used to consider its complex structure and frequently used tasks. Main findings in related PUI studies were summarized based on this framework. Our secondary aim was to compare compulsivity-related features between different PUI subtypes. Thirty-four empirical studies were retained, comprising 41 task-results and 35 independent data sets. Overall, individuals with PUI showed more consistent deficits in attentional biases and were relatively intact in set-shifting. Few studies have examined cognitive flexibility and habit learning, and more evidence is thus needed to establish reliable conclusions. Moreover, most studies focused on internet gaming disorder, whereas other PUI sub-types were not sufficiently examined. This systematic review highlights the use of the four-domain framework for advancing understanding of mechanisms underlying compulsivity in PUI. Related therapeutic implications and future directions are discussed.

Epistatic Features and Machine Learning Improve Alzheimer's Disease Risk Prediction Over Polygenic Risk Scores.

Polygenic risk scores (PRS) are linear combinations of genetic markers weighted by effect size that are commonly used to predict disease risk. For complex heritable diseases such as late-onset Alzheimer's disease (LOAD), PRS models fail to capture much of the heritability. Additionally, PRS models are highly dependent on the population structure of the data on which effect sizes are assessed and have poor generalizability to new data. The goal of this study is to construct a paragenic risk score that, in addition to single genetic marker data used in PRS, incorporates epistatic interaction features and machine learning methods to predict risk for LOAD. We construct a new state-of-the-art genetic model for risk of Alzheimer's disease. Our approach innovates over PRS models in two ways: First, by directly incorporating epistatic interactions between SNP loci using an evolutionary algorithm guided by shared pathway information; and second, by estimating risk via an ensemble of non-linear machine learning models rather than a single linear model. We compare the paragenic model to several PRS models from the literature trained on the same dataset. The paragenic model is significantly more accurate than the PRS models under 10-fold cross-validation, obtaining an AUC of 83% and near-clinically significant matched sensitivity/specificity of 75%. It remains significantly more accurate when evaluated on an independent holdout dataset and maintains accuracy within APOE genotype strata. Paragenic models show potential for improving disease risk prediction for complex heritable diseases such as LOAD over PRS models.

Identification of Family-Specific Features in Cas9 and Cas12 Proteins: A Machine Learning Approach Using Complete Protein Feature Spectrum.

The recent development of CRISPR-Cas technology holds promise to correct gene-level defects for genetic diseases. The key element of the CRISPR-Cas system is the Cas protein, a nuclease that can edit the gene of interest assisted by guide RNA. However, these Cas proteins suffer from inherent limitations such as large size, low cleavage efficiency, and off-target effects, hindering their widespread application as a gene editing tool. Therefore, there is a need to identify novel Cas proteins with improved editing properties, for which it is necessary to understand the underlying features governing the Cas families. In this study, we aim to elucidate the unique protein features associated with Cas9 and Cas12 families and identify the features distinguishing each family from non-Cas proteins. Here, we built Random Forest (RF) binary classifiers to distinguish Cas12 and Cas9 proteins from non-Cas proteins, respectively, using the complete protein feature spectrum (13,494 features) encoding various physiochemical, topological, constitutional, and coevolutionary information on Cas proteins. Furthermore, we built multiclass RF classifiers differentiating Cas9, Cas12, and non-Cas proteins. All the models were evaluated rigorously on the test and independent data sets. The Cas12 and Cas9 binary models achieved a high overall accuracy of 92% and 95% on their respective independent data sets, while the multiclass classifier achieved an F1 score of close to 0.98. We observed that Quasi-Sequence-Order (QSO) descriptors like Schneider.lag and Composition descriptors like charge, volume, and polarizability are predominant in the Cas12 family. Conversely Amino Acid Composition descriptors, especially Tripeptide Composition (TPC), predominate the Cas9 family. Four of the top 10 descriptors identified in Cas9 classification are tripeptides PWN, PYY, HHA, and DHI, which are seen to be conserved across all Cas9 proteins and located within different catalytically important domains of the Streptococcus pyogenes Cas9 (SpCas9) structure. Among these, DHI and HHA are well-known to be involved in the DNA cleavage activity of the SpCas9 protein. Mutation studies have highlighted the significance of the PWN tripeptide in PAM recognition and DNA cleavage activity of SpCas9, while Y450 from the PYY tripeptide plays a crucial role in reducing off-target effects and improving the specificity in SpCas9. Leveraging our machine learning (ML) pipeline, we identified numerous Cas9 and Cas12 family-specific features. These features offer valuable insights for future experimental and computational studies aiming at designing Cas systems with enhanced gene-editing properties. These features suggest plausible structural modifications that can effectively guide the development of Cas proteins with improved editing capabilities.

Transcriptome Study of 2 BlackCohorts Reveals cis Long Noncoding RNAs Associated With Hypertension-Related mRNAs.

Long noncoding RNAs (lncRNAs) have emerged as critical regulators of the expression of genes involved in cardiovascular diseases. This project aims to identify circulating lncRNAs associated with protein-coding mRNAs differentially expressed between hypertensive and normotensive individuals and establish their link with hypertension. The analyses were conducted in 3 main steps: (1) an unbiased whole blood transcriptome-wide analysis was conducted to identify and replicate protein-coding genes differentially expressed by hypertension status in 497 and 179 Black individuals from the GENE-FORECAST (Genomics, Environmental Factors and the Social Determinants of Cardiovascular Disease in African-Americans Study) and MH-GRID (Minority Health Genomics and Translational Research Bio-Repository Database) studies, respectively. Subsequently, (2) proximal lncRNAs, termed cis lncRNA quantitative trait loci, associated with each mRNA were identified in the GENE-FORECAST study and replicated in the MH-GRID study. Finally, (3) the lncRNA quantitative trait loci were used as predictors in a random forest model to predict hypertension in both data sets. A total of 129 mRNAs were significantly differentially expressed between normotensive and hypertensive individuals in both data sets. The lncRNA-mRNA association analysis revealed 249 cis lncRNA quantitative trait loci associated with 102 mRNAs, including VAMP2 (vesicle-associated membrane protein 2), mitogen-activated protein kinase kinase 3, CCAAT enhancer binding protein beta, and lymphocyte antigen 6 complex, locus E. The 249 lncRNA quantitative trait loci predicted hypertension with an area under the curve of 0.79 and 0.71 in GENE-FORECAST and MH-GRID studies, respectively. This study leveraged a significant sample of Black individuals, a population facing a disproportionate burden of hypertension. The analyses unveiled a total of 271 lncRNA-mRNA relationships involving mRNAs that play critical roles in vascular pathways relevant to blood pressure regulation. The compelling findings, consistent across 2 independent data sets, establish a reliable foundation for designing invitro/in vivo experiments.

HeteroTCR: A heterogeneous graph neural network-based method for predicting peptide-TCR interaction

Identifying interactions between T-cell receptors (TCRs) and immunogenic peptides holds profound implications across diverse research domains and clinical scenarios. Unsupervised clustering models (UCMs) cannot predict peptide-TCR binding directly, while supervised predictive models (SPMs) often face challenges in identifying antigens previously unencountered by the immune system or possessing limited TCR binding repertoires. Therefore, we propose HeteroTCR, an SPM based on Heterogeneous Graph Neural Network (GNN), to accurately predict peptide-TCR binding probabilities. HeteroTCR captures within-type (TCR-TCR or peptide-peptide) similarity information and between-type (peptide-TCR) interaction insights for predictions on unseen peptides and TCRs, surpassing limitations of existing SPMs. Our evaluation shows HeteroTCR outperforms state-of-the-art models on independent datasets. Ablation studies and visual interpretation underscore the Heterogeneous GNN module’s critical role in enhancing HeteroTCR’s performance by capturing pivotal binding process features. We further demonstrate the robustness and reliability of HeteroTCR through validation using single-cell datasets, aligning with the expectation that pMHC-TCR complexes with higher predicted binding probabilities correspond to increased binding fractions.

Advances in Inflammatory Bowel Disease Diagnostics: Machine Learning and Genomic Profiling Reveal Key Biomarkers for Early Detection.

This study, utilizing high-throughput technologies and Machine Learning (ML), has identified gene biomarkers and molecular signatures in Inflammatory Bowel Disease (IBD). We could identify significant upregulated or downregulated genes in IBD patients by comparing gene expression levels in colonic specimens from 172 IBD patients and 22 healthy individuals using the GSE75214 microarray dataset. Our ML techniques and feature selection methods revealed six Differentially Expressed Gene (DEG) biomarkers (VWF, IL1RL1, DENND2B, MMP14, NAAA, and PANK1) with strong diagnostic potential for IBD. The Random Forest (RF) model demonstrated exceptional performance, with accuracy, F1-score, and AUC values exceeding 0.98. Our findings were rigorously validated with independent datasets (GSE36807 and GSE10616), further bolstering their credibility and showing favorable performance metrics (accuracy: 0.841, F1-score: 0.734, AUC: 0.887). Our functional annotation and pathway enrichment analysis provided insights into crucial pathways associated with these dysregulated genes. DENND2B and PANK1 were identified as novel IBD biomarkers, advancing our understanding of the disease. The validation in independent cohorts enhances the reliability of these findings and underscores their potential for early detection and personalized treatment of IBD. Further exploration of these genes is necessary to fully comprehend their roles in IBD pathogenesis and develop improved diagnostic tools and therapies. This study significantly contributes to IBD research with valuable insights, potentially greatly enhancing patient care.

Convolutional neural networks to classify human stress that occurs during in‐field sugarcane harvesting: A case study

AbstractAssessing human stress in agriculture proves to be a complex and time‐intensive endeavor within the field of ergonomics, particularly for the development of agricultural systems. This methodology involves the utilization of instrumentation and the establishment of a dedicated laboratory setup. The complexity arises from the need to capture and analyze various physiological and psychological indicators, such as heart rate (HR), muscle activity, and subjective feedback to comprehensively assess the impact of farm operations on subjects. The instrumentation typically includes wearable devices, sensors, and monitoring equipment to gather real‐time data of subject during the performance of farm operations. Deep learning (DL) models currently achieve human performance levels on real‐world face recognition tasks. In this study, we went beyond face recognition and experimented with the recognition of human stress based on facial features during the drudgery‐prone agricultural operation of sugarcane harvesting. This is the first research study for deploying artificial intelligence‐driven DL techniques to identify human stress in agriculture instead of monitoring several ergonomic characteristics. A total of 20 (10 each for male and female) subjects comprising 4300 augmented RGB images (215 per subject) were acquired during sugarcane harvesting seasons and then these images were deployed for training (80%) and validation (20%). Human stress and nonstress states were determined based on four ergonomic physiological parameters: heart rate (ΔHR), oxygen consumption rate (OCR), energy expenditure rate (EER), and acceptable workload (AWL). Stress was defined when ΔHR, OCR, EER, and AWL reached or exceeded certain standard threshold values. Four convolutional neural network‐based DL models (1) DarkNet53, (2) InceptionV3, (3) MobileNetV2 and (4) ResNet50 were selected due to their remarkable feature extraction abilities, simple and effective implementation to edge computation devices. In all four DL models, training performance results delivered training accuracy ranging from 73.8% to 99.1% at combinations of two mini‐batch sizes and four levels of epochs. The maximum training accuracies were 99.1%, 99.0%, 97.7%, and 95.4% at the combination of mini‐batch size 16 and 25 epochs for DarkNet53, InceptionV3, ResNet50, and MobileNetV2, respectively. Due to the best performance, DarkNet53 was tested further on an independent data set of 100 images and found 89.8%–93.3% confident to classify stressed images for female subjects while 92.2%–94.5% for male subjects, though it was trained on the integrated data set. The comparative classification of the developed model and ergonomic measurements for stress classification was carried out with a net accuracy of 88% where there were few instances of wrong classifications.

New Model and Public Online Prediction Platform for Risk Stratification of Vocal Cord Leukoplakia.

To extract texture features from vocal cord leukoplakia (VCL) images and establish a VCL risk stratification prediction model using machine learning (ML) techniques. A total of 462 patients with pathologically confirmed VCL were retrospectively collected and divided into low-risk and high-risk groups. We use a 5-fold cross validation method to ensure the generalization ability of the model built using the included dataset and avoid overfitting. Totally 504 texture features were extracted from each laryngoscope image. After feature selection, 10 ML classifiers were utilized to construct the model. The SHapley Additive exPlanations (SHAP) was employed for feature analysis. To evaluate the model, accuracy, sensitivity, specificity, and the area under the receiver operating characteristic (ROC) curve (AUC) were utilized. In addition, the model was transformed into an online application for public use and further tested in an independent dataset with 52 cases of VCL. A total of 12 features were finally selected, random forest (RF) achieved the best model performance, the mean accuracy, sensitivity, specificity, and AUC of the 5-fold cross validation were 92.2 ± 4.1%, 95.6 ± 4.0%, 85.8 ± 5.8%, and 90.7 ± 4.9%, respectively. The result is much higher than the clinicians (AUC between 63.1% and 75.2%). The SHAP algorithm ranks the importance of 12 texture features to the model. The test results of the additional independent datasets were 92.3%, 95.7%, 90.0%, and 93.3%, respectively. The proposed VCL risk stratification prediction model, which has been developed into a public online prediction platform, may be applied in practical clinical work. 3 Laryngoscope, 2024.

Identification of RNA‐dependent liquid‐liquid phase separation proteins using an artificial intelligence strategy

RNA-dependent liquid-liquid phase separation (LLPS) proteins play critical roles in cellular processes such as stress granule formation, DNA repair, RNA metabolism, germ cell development, and protein translation regulation. The abnormal behavior of these proteins is associated with various diseases, particularly neurodegenerative disorders like amyotrophic lateral sclerosis and frontotemporal dementia, making their identification crucial. However, conventional biochemistry-based methods for identifying these proteins are time-consuming and costly. Addressing this challenge, our study developed a robust computational model for their identification. We constructed a comprehensive dataset containing 137 RNA-dependent and 606 non-RNA-dependent LLPS protein sequences, which were then encoded using amino acid composition, composition of K-spaced amino acid pairs, Geary autocorrelation, and conjoined triad methods. Through a combination of correlation analysis, mutual information scoring, and incremental feature selection, we identified an optimal feature subset. This subset was used to train a random forest model, which achieved an accuracy of 90% when tested against an independent dataset. This study demonstrates the potential of computational methods as efficient alternatives for the identification of RNA-dependent LLPS proteins. To enhance the accessibility of the model, a user-centric web server has been established and can be accessed via the link: http://rpp.lin-group.cn.

Integrative bioinformatics approach yields a novel gene expression risk model for prognosis and progression prediction in prostate cancer.

Prostate cancer (PCa), a prevalent malignancy among elderly males, exhibits a notable rate of advancement, even when subjected to conventional androgen deprivation therapy or chemotherapy. An effective progression prediction model would prove invaluable in identifying patients with a higher progression risk. Using bioinformatics strategies, we integrated diverse data sets of PCa to construct a novel risk model predicated on gene expression and progression-free survival (PFS). The accuracy of the model was assessed through validation using an independent data set. Eight genes were discerned as independent prognostic factors and included in the prediction model. Patients assigned to the high-risk cohort demonstrated a diminished PFS, and the areas under the curve of our model in the validation set for 1-year, 3-year, and 5-year PFS were 0.9325, 0.9041 and 0.9070, respectively. Additionally, through the application of single-cell RNA sequencing to two castration-related prostate cancer (CRPC) samples and two hormone-related prostate cancer (HSPC) samples, we discovered that luminal cells within CRPC exhibited an elevated risk score. Subsequent molecular biology experiments corroborated our findings, illustrating heightened SYK expression levels within tumour tissues and its contribution to cancer cell migration. We found that the knockdown of SYK could inhibit migration in PCa cells. Our progression-related risk model demonstrated the potential prognostic value of SYK and indicated its potential as a target for future diagnosis and treatment strategies in PCa management.

Prediction of Plant Ubiquitylation Proteins and Sites by Fusing Multiple Features

Introduction: Protein ubiquitylation is an important post-translational modification (PTM), which is considered to be one of the most important processes regulating cell function and various diseases. Therefore, accurate prediction of ubiquitylation proteins and their PTM sites is of great significance for the study of basic biological processes and the development of related drugs. Researchers have developed some large-scale computational methods to predict ubiquitylation sites, but there is still much room for improvement. Much of the research related to ubiquitylation is cross-species while the life pattern is diversified, and the prediction method always shows its specificity in practical application. This study just aims at the issue of plants and has constructed computational methods for identifying ubiquitylation protein and ubiquitylation sites. Method: In this work, we constructed two predictive models to identify plant ubiquitylation proteins and sites. First, in the ubiquitylation proteins prediction model, in order to better reflect protein sequence information and obtain better prediction results, the KNN scoring matrix model based on functional domain Gene Ontology (GO) annotation and word embedding model, i.e. Skip-Gram and Continuous Bag of Words (CBOW), are used to extract the features, and the light gradient boosting machine (LGBM) is selected as the ubiquitylation proteins prediction engine. Results: As a result, accuracy (ACC), Precision, recall rate (Recall), F1_score and AUC are respectively 85.12%, 80.96%, 72.80%, 76.37% and 0.9193 in the 10-fold cross-validations on independent dataset. In the ubiquitylation sites prediction model, Skip-Gram, CBOW and enhanced amino acid composition (EAAC) feature extraction codes were used to extract protein sequence fragment features, and the predicted results on training and independent test data have also achieved good performance. Conclusion: In a word, the comparison results demonstrate that our models have a decided advantage in predicting ubiquitylation proteins and sites, and it may provide useful insights for studying the mechanisms and modulation of ubiquitination pathways

A liquid biopsy-based assay for classifying breast cancer from benign nodule.

e12572 Background: Breast cancer (BRCA) ranks as the most frequently diagnosed non-skin cancer and stands as the second leading cause of cancer-related mortality in women in the United States. Globally, 2.3 million women were diagnosed with breast cancer in 2020, resulting in 685,000 deaths. Therefore, accurate classification of breast cancer could contribute to immediate treatment and improve outcomes for individuals diagnosed with breast cancer. Our study introduces an ensemble model utilizing the distinctive characteristics of cell-free DNA (cfDNA) fragmentation from whole genome sequencing (WGS) for the classification of BRCA and benign nodule. Methods: The study cohort consisted of 214 individuals diagnosed with breast cancer (BRCA) and 211 individuals with benign breast nodule. Employing whole-genome sequencing (WGS), we subjected plasma samples from all participants to comprehensive fragmentomic profiling. Three fragmentomic features: copy number variation (CNV), fragmentation size coverage (FSC), and fragmentation size distribution (FSD), were utilized by machine learning algorithms. An ensemble model integrated results from each algorithm then generated a cancer score ranging from 0 to 1 for each sample. The model was trained exclusively using the training cohort and the model’s performance was assessed in the independent validation cohort. Results: Our ensemble model demonstrated a sensitivity of 80.2% and a specificity of 95.3% in training cohort, with an AUC of 0.966. In the validation cohort, the model maintained its effectiveness, with a sensitivity of 73.6%, a specificity of 89.3% supported by an AUC of 0.919. Notably, the model showed various predictive ability across different stages of breast cancer. In the training cohort, it exhibited a sensitivity of 76.4% for Stage I, 81.8% for Stage II, and 88.9% for Stage III, indicating superior predictive power at advanced progression levels. Additionally, the model demonstrated a consistent sensitivity of 75% for Stage I in the validation cohort, affirming its reliable performance across independent datasets. When stratified by molecular subtypes, the model showed relatively lower scores and sensitivity BRCA with HR+/HER2+, compared to HR-/HER2+ and HR+/HER2- subgroups, as well as Triple-Negative Breast Cancer. Finally, the model generated unbiased predictions across various histology subtypes, accurately discerning ductal carcinoma and other BRCA histology subtypes in training and validation cohorts. Conclusions: The ensemble model that leverages non-invasive liquid biopsy-based assay for the accurate classification of BRCA and benign nodule provides a critical opportunity for effective intervention for cancer patients. This approach, facilitating swift and appropriate treatments, and optimizing the allocation of healthcare resources, underscores its significance in the ongoing pursuit of enhancing breast cancer care and outcomes.

A comparison of wavelet-based action potential detection from the NeuroAmp and the Iowa Bioengineering Nerve Traffic Analysis system.

Microneurographic recordings of muscle sympathetic nerve activity (MSNA) reflect postganglionic sympathetic axonal activity directed toward the skeletal muscle vasculature. Recordings are typically evaluated for spontaneous bursts of MSNA; however, the filtering and integration of raw neurograms to obtain multiunit bursts conceals the underlying c-fiber discharge behavior. The continuous wavelet transform with matched mother wavelet has permitted the assessment of action potential discharge patterns, but this approach uses a mother wavelet optimized for an amplifier that is no longer commercially available (University of Iowa Bioengineering Nerve Traffic Analysis System; Iowa NTA). The aim of this project was to determine the morphology and action potential detection performance of mother wavelets created from the commercially available NeuroAmp (ADinstruments), from distinct laboratories, compared with a mother wavelet generated from the Iowa NTA. Four optimized mother wavelets were generated in a two-phase iterative process from independent datasets, collected by separate laboratories (one Iowa NTA, three NeuroAmp). Action potential extraction performance of each mother wavelet was compared for each of the NeuroAmp-based datasets. The total number of detected action potentials was not significantly different across wavelets. However, the predictive value of action potential detection was reduced when the Iowa NTA wavelet was used to detect action potentials in NeuroAmp data, but not different across NeuroAmp wavelets. To standardize approaches, we recommend a NeuroAmp-optimized mother wavelet be used for the evaluation of sympathetic action potential discharge behavior when microneurographic data are collected with this system.NEW & NOTEWORTHY The morphology of custom mother wavelets produced across laboratories using the NeuroAmp was highly similar, but distinct from the University of Iowa Bioengineering Nerve Traffic Analysis System. Although the number of action potentials detected was similar between collection systems and mother wavelets, the predictive value differed. Our data suggest action potential analysis using the continuous wavelet transform requires a mother wavelet optimized for the collection system.

Adaptive Multimodel Knowledge Transfer Matrix Machine for EEG Classification.

The emerging matrix learning methods have achieved promising performances in electroencephalogram (EEG) classification by exploiting the structural information between the columns or rows of feature matrices. Due to the intersubject variability of EEG data, these methods generally need to collect a large amount of labeled individual EEG data, which would cause fatigue and inconvenience to the subjects. Insufficient subject-specific EEG data will weaken the generalization capability of the matrix learning methods in neural pattern decoding. To overcome this dilemma, we propose an adaptive multimodel knowledge transfer matrix machine (AMK-TMM), which can selectively leverage model knowledge from multiple source subjects and capture the structural information of the corresponding EEG feature matrices. Specifically, by incorporating least-squares (LS) loss with spectral elastic net regularization, we first present an LS support matrix machine (LS-SMM) to model the EEG feature matrices. To boost the generalization capability of LS-SMM in scenarios with limited EEG data, we then propose a multimodel adaption method, which can adaptively choose multiple correlated source model knowledge with a leave-one-out cross-validation strategy on the available target training data. We extensively evaluate our method on three independent EEG datasets. Experimental results demonstrate that our method achieves promising performances on EEG classification.

Dynamic Evolution of Infarct Volumes at MRI in Ischemic Stroke Due to Large Vessel Occlusion.

The typical infarct volume trajectories in stroke patients, categorized as slow or fast progressors, remain largely unknown. This study aimed to reveal the characteristic spatiotemporal evolutions of infarct volumes caused by large vessel occlusion (LVO) and show that such growth charts help anticipate clinical outcomes. We conducted a secondary analysis from prospectively collected databases (FRAME, 2017-2019; ETIS, 2015-2022). We selected acute MRI data from anterior LVO stroke patients with witnessed onset, which were divided into training and independent validation datasets. In the training dataset, using Gaussian mixture analysis, we classified the patients into 3 growth groups based on their rate of infarct growth (diffusion volume/time-to-imaging). Subsequently, we extrapolated pseudo-longitudinal models of infarct growth for each group and generated sequential frequency maps to highlight the spatial distribution of infarct growth. We used these charts to attribute a growth group to the independent patients from the validation dataset. We compared their 3-month modified Rankin scale (mRS) with the predicted values based on a multivariable regression model from the training dataset that used growth group as an independent variable. We included 804 patients (median age 73.0 years [interquartile range 61.2-82.0 years]; 409 men). The training dataset revealed nonsupervised clustering into 11% (74/703) slow, 62% (437/703) intermediate, and 27% (192/703) fast progressors. Infarct volume evolutions were best fitted with a linear (r = 0.809; p < 0.001), cubic (r = 0.471; p < 0.001), and power (r = 0.63; p < 0.001) function for the slow, intermediate, and fast progressors, respectively. Notably, the deep nuclei and insular cortex were rapidly affected in the intermediate and fast groups with further cortical involvement in the fast group. The variable growth group significantly predicted the 3-month mRS (multivariate odds ratio 0.51; 95% CI 0.37-0.72, p < 0.0001) in the training dataset, yielding a mean area under the receiver operating characteristic curve of 0.78 (95% CI 0.66-0.88) in the independent validation dataset. We revealed spatiotemporal archetype dynamic evolutions following LVO stroke according to 3 growth phenotypes called slow, intermediate, and fast progressors, providing insight into anticipating clinical outcome. We expect this could help in designing neuroprotective trials aiming at modulating infarct growth before EVT.

A multi-ethnic reference panel to impute HLA classical and non-classical class I alleles in admixed samples: Testing imputation accuracy in an admixed sample from Brazil.

The MHC class I region contains crucial genes for the innate and adaptive immune response, playing a key role in susceptibility to many autoimmune and infectious diseases. Genome-wide association studies have identified numerous disease-associated SNPs within this region. However, these associations do not fully capture the immune-biological relevance of specific HLA alleles. HLA imputation techniques may leverage available SNP arrays by predicting allele genotypes based on the linkage disequilibrium between SNPs and specific HLA alleles. Successful imputation requires diverse and large reference panels, especially for admixed populations. This study employed a bioinformatics approach to call SNPs and HLA alleles in multi-ethnic samples from the 1000 genomes (1KG) dataset and admixed individuals from Brazil (SABE), utilising 30X whole-genome sequencing data. Using HIBAG, we created three reference panels: 1KG (n = 2504), SABE (n = 1171), and the full model (n = 3675) encompassing all samples. In extensive cross-validation of these reference panels, the multi-ethnic 1KG reference exhibited overall superior performance than the reference with only Brazilian samples. However, the best results were achieved with the full model. Additionally, we expanded the scope of imputation by developing reference panels for non-classical, MICA, MICB and HLA-H genes, previously unavailable for multi-ethnic populations. Validation in an independent Brazilian dataset showcased the superiority of our reference panels over the Michigan Imputation Server, particularly in predicting HLA-B alleles among Brazilians. Our investigations underscored the need to enhance or adapt reference panels to encompass the target population's genetic diversity, emphasising the significance of multiethnic references for accurate imputation across different populations.

Real-world and clinical trial validation of a deep learning radiomic biomarker for PD-(L)1 immune checkpoint inhibitor response in stage IV NSCLC.

102 Background: Immune checkpoint inhibitor (ICI) therapy is standard-of-care for treatment of mutation-negative advanced non-small cell lung cancer (NSCLC). However, given known inaccuracy of PD-(L)1 markers, there is an unmet need to better identify patients most likely to derive clinical benefit from ICI. We developed and externally validated a generalizable CT imaging-based biomarker to predict response to ICI. Methods: We developed and validated a deep learning radiomic biomarker using an internally curated real-world dataset (RWD) of 2,010 stage IV NSCLC patients treated with PD-(L)1 ICIs in academic and community settings from US and Europe. Patients with missing baseline imaging, missing follow-up data, or EGFR/ALK oncogenic driver mutations were excluded, resulting in a total of 1,188 subjects. This RWD consisted of a discovery cohort (Dataset A, N=844) and a temporally distinct holdout cohort (Dataset B, N=344), which were used to generate performance metrics of the biomarker. To test generalizability, we validated our biomarker in a prospective clinical trial dataset evaluating Sasanlimab in PD-(L)1 therapy-naïve, advanced NSCLC patients (NCT02573259, Dataset C, N=54). We utilized a two-stage learning approach to model 6-month PFS. First, we used an independent multi-task deep-learning feature extractor trained on 19,184 whole chest CT scans. Second, we input the extracted features into a Cox proportional hazard (CoxPH) model along with age, sex, and baseline lesion measurements (sum of longest diameter and distant metastases counts), and the model generated a time-dependent PFS function and a response score. We performed 6-fold cross-validation on Dataset A to train and evaluate the models, which were subsequently ensembled and applied to independent Datasets B and C. To assess independence from PD-L1 status and key demographic covariates, we herein report multivariate adjusted hazard ratios (HR) for the group identified as low-risk based on the biomarker. Results: In Dataset A, the biomarker showed a cross-validation PFS adjusted HR of 0.49 (95% CI 0.38-0.63) in the all-comers cohort and 0.28 (0.17-0.46) in the first-line ICI monotherapy cohort (1LMono). In Dataset B, the PFS adjusted HRs were 0.54 (0.35-0.83) in all-comers and 0.18 (0.05-0.61) in 1LMono. In Dataset C, the adjusted HRs were 0.30 (0.14-0.68) for PFS, 0.29 (0.10-0.83) for OS, 0.31 (0.14-0.72) for TTP. Conclusions: In our validations in RWD and clinical trial cohorts, a deep-learning radiomic biomarker based on routine pre-treatment CT scans predicted response to ICI and stratified patients independently from PD-L1 status. This tool may inform clinical decision-making, such as to help guide whether concomitant chemotherapy may not be needed. In future work, we plan to further validate our approach in larger prospective datasets and expand its use to new indications.

What can be learnt from UHECR anisotropies observations

Context. Various signals of anisotropy of the ultra-high-energy cosmic rays (UHECRs) have recently been reported, whether at large angular scales, with a dipole modulation in right ascension observed in the data of the Pierre Auger observatory (Auger), as discussed in the first paper accompanying the present one, or at intermediate angular scales, with flux excesses identified in specific directions by Auger and the Telescope Array (TA) Collaborations. Aims. We investigated the implications of the current data regarding these intermediate scale anisotropies, and examined to what extent they can be used to shed light on the origin of UHECRs, and constrain the astrophysical and/or physical parameters of the viable source scenarios. We also investigated what could be learnt from the study of the evolution of the various UHECR anisotropy signals, and discussed the expected benefit of an increased exposure of the UHECR sky using future observatories. Methods. We simulated realistic UHECR sky maps for a wide range of astrophysical scenarios satisfying the current observational constraints, with the assumption that the UHECR source distribution follows that of the galaxies in the Universe, also implementing possible biases towards specific classes of sources. In each case, several scenarios were explored with different UHECR source compositions and spectra, a range of source densities and different models of the Galactic magnetic field. We also implemented the Auger sky coverage, and explored various levels of statistics. For each scenario, we produced 300 independent datasets on which we applied similar analyses as those recently used by the Auger Collaboration, searching for flux excesses through either blind or targeted searches and quantifying correlations with predefined source catalogues through a likelihood analysis. Results. We find the following. First, with reasonable choices of the parameters, the investigated astrophysical scenarios can easily account for the significance of the anisotropies reported by Auger, even with large source densities. Second, the direction in which the maximum flux excess is found in the Auger data differs from the region where it is found in most of our simulated datasets, although an angular distance as large as that between the Auger direction and the direction expected from the simulated models at infinite statistics, of the order of ∼20°, occurs in ∼25% of the cases. Third, for datasets simulated with the same underlying astrophysical scenario, and thus the same actual UHECR sources, the significance with which the isotropy hypothesis is rejected through the Auger likelihood analysis can be largest either when ‘all galaxies’ or when only ‘starburst’ galaxies are used to model the signal, depending on which model is used to model the Galactic magnetic field and the resulting deflections. Fourth, the study of the energy evolution of the anisotropy patterns can be very instructive and provide new astrophysical insight about the origin of the UHECRs. Fifth, the direction in which the most significant flux excess is found in the Auger dataset above 8 EeV appears to essentially disappear in the dataset above 32 EeV, and, conversely, the maximum excess at high energy has a much reduced significance in the lower energy dataset. Sixth, both of these appear to be very uncommon in the simulated datasets, which could point to a failure of some generic assumption in the investigated astrophysical scenarios, such as the dominance of one type of source with essentially the same composition and spectrum in the observed UHECR flux above the ankle. Seventh, given the currently observed level of anisotropy signals, a meaningful measurement of their energy evolution, say from 10 EeV to the highest energies, will require a significant increase in statistics and a new generation of UHECR observatories.