Independent Test Set Research Articles

Exploring the Combination of Computer Vision and Surgical Neuroanatomy: A Workflow Involving Artificial Intelligence for the Identification of Skull Base Foramina

The skull base is a complex region in neurosurgery, featuring numerous foramina. Accurate identification of these foramina is imperative to avoid intraoperative complications and to facilitate educational progress in neurosurgical trainees. The intricate landscape of the skull base often challenges both clinicians and learners, necessitating innovative identification solutions. We aimed to develop a computer vision model that automates the identification and labeling of the skull base foramina from various image formats, enhancing surgical planning and educational outcomes. We employed a deep learning methodology, specifically using a convolutional neural network architecture. Our model was trained on a dataset comprising of 3560 high-resolution, annotated images of the skull base, taken from various perspectives and lighting conditions to ensure model generalizability. Model performance was quantitatively assessed using precision and recall metrics. The convolutional neural network model demonstrated strong performance, achieving an average precision of 0.77. At a confidence threshold of 0.28, the model reached an optimal precision of 90.4% and a recall of 89.6%. Validation on an independent test set of images corroborated the model's capability to consistently and accurately identify and label multiple skull base foramina across diverse imaging scenarios. This study successfully introduces a highly accurate computer vision model tailored for the identification of skull base foramina, illustrating the model's potential as a transformative tool in anatomical education and intraoperative structure visualization. The findings suggest promising avenues for future research into automated anatomical recognition models, suggesting a trajectory toward increasingly sophisticated aids in neurosurgical operations and education.

Automatic Transcranial Sonography-Based Classification of Parkinson's Disease Using a Novel Dual-Channel CNXV2-DANet.

Transcranial sonography (TCS) has been introduced to assess hyper-echogenicity in the substantia nigra of the midbrain for Parkinson's disease (PD); however, its subjective and resource-demanding nature has impeded its widespread application. An AI-empowered TCS-based PD classification tool is greatly demanding, yet relevant research is severely scarce. Therefore, we proposed a novel dual-channel CNXV2-DANet for TCS-based PD classification using a large cohort. A total of 1176 TCS images from 588 subjects were retrospectively enrolled from Beijing Tiantan Hospital, encompassing both the left and right side of the midbrain for each subject. The entire dataset was divided into a training/validation/testing set at a ratio of 70%/15%/15%. Development of the proposed CNXV2-DANet was performed on the training set with comparisons between the single-channel and dual-channel input settings; model evaluation was conducted on the independent testing set. The proposed dual-channel CNXV2-DANet was compared against three state-of-the-art networks (ConvNeXtV2, ConvNeXt, Swin Transformer). The results demonstrated that both CNXV2-DANet and ConvNeXt V2 performed more superiorly under dual-channel inputs than the single-channel input. The dual-channel CNXV2-DANet outperformed the single-channel, achieving superior average metrics for accuracy (0.839 ± 0.028), precision (0.849 ± 0.014), recall (0.845 ± 0.043), F1-score (0.820 ± 0.038), and AUC (0.906 ± 0.013) compared with the single channel metrics for accuracy (0.784 ± 0.037), precision (0.817 ± 0.090), recall (0.748 ± 0.093), F1-score (0.773 ± 0.037), and AUC (0.861 ± 0.047). Furthermore, the dual-channel CNXV2-DANet outperformed all other networks (all p-values < 0.001). These findings suggest that the proposed dual-channel CNXV2-DANet may provide the community with an AI-empowered TCS-based tool for PD assessment.

Multitask Learning on Graph Convolutional Residual Neural Networks for Screening of Multitarget Anticancer Compounds.

Recently, various modern experimental screening pipelines and assays have been developed to find promising anticancer drug candidates. However, it is time-consuming and almost infeasible to screen an immense number of compounds for anticancer activity via experimental approaches. To partially address this issue, several computational advances have been proposed. In this study, we present iACP-GCR, a model based on multitask learning on graph convolutional residual neural networks with two types of shortcut connections, to identify multitarget anticancer compounds. In our architecture, the graph convolutional residual neural networks are shared by all the prediction tasks before being separately customized. The NCI-60 data set, one of the most reliable and well-known sources of experimentally verified compounds, was used to develop our model. From that data set, we collected and refined data about compounds screened across nine cancer types (panels), including breast, central nervous system, colon, leukemia, nonsmall cell lung, melanoma, ovarian, prostate, and renal, for model training and evaluation. The model performance evaluated on an independent test set shows that iACP-GCR surpasses the three advanced computational methods for multitask learning. The integration of two shortcut connection types in the shared networks also improves the prediction efficiency. We also deployed the model as a public web server to assist the research community in screening potential anticancer compounds.

A Point Cloud Graph Neural Network for Protein-Ligand Binding Site Prediction.

Predicting protein-ligand binding sites is an integral part of structural biology and drug design. A comprehensive understanding of these binding sites is essential for advancing drug innovation, elucidating mechanisms of biological function, and exploring the nature of disease. However, accurately identifying protein-ligand binding sites remains a challenging task. To address this, we propose PGpocket, a geometric deep learning-based framework to improve protein-ligand binding site prediction. Initially, the protein surface is converted into a point cloud, and then the geometric and chemical properties of each point are calculated. Subsequently, the point cloud graph is constructed based on the inter-point distances, and the point cloud graph neural network (GNN) is applied to extract and analyze the protein surface information to predict potential binding sites. PGpocket is trained on the scPDB dataset, and its performance is verified on two independent test sets, Coach420 and HOLO4K. The results show that PGpocket achieves a 58% success rate on the Coach420 dataset and a 56% success rate on the HOLO4K dataset. These results surpass competing algorithms, demonstrating PGpocket's advancement and practicality for protein-ligand binding site prediction.

Expert-Level Detection of Referable Glaucoma from Fundus Photographs in a Safety Net Population: The AI and Teleophthalmology in Los Angeles Initiative.

To develop and test a deep learning (DL) algorithm for detecting referable glaucoma in the Los Angeles County (LAC) Department of Health Services (DHS) teleretinal screening program. Fundus photographs and patient-level labels of referable glaucoma (defined as cup-to-disc ratio [CDR] ≥ 0.6) provided by 21 trained optometrist graders were obtained from the LAC DHS teleretinal screening program. A DL algorithm based on the VGG-19 architecture was trained using patient-level labels generalized to images from both eyes. Area under the receiver operating curve (AUC), sensitivity, and specificity were calculated to assess algorithm performance using an independent test set that was also graded by 13 clinicians with one to 15 years of experience. Algorithm performance was tested using reference labels provided by either LAC DHS optometrists or an expert panel of 3 glaucoma specialists. 12,098 images from 5,616 patients (2,086 referable glaucoma, 3,530 non-glaucoma) were used to train the DL algorithm. In this dataset, mean age was 56.8 ± 10.5 years with 54.8% females and 68.2% Latinos, 8.9% Blacks, 2.7% Caucasians, and 6.0% Asians. 1,000 images from 500 patients (250 referable glaucoma, 250 non-glaucoma) with similar demographics (p ≥ 0.57) were used to test the DL algorithm. Algorithm performance matched or exceeded that of all independent clinician graders in detecting patient-level referable glaucoma based on LAC DHS optometrist (AUC = 0.92) or expert panel (AUC = 0.93) reference labels. Clinician grader sensitivity (range: 0.33-0.99) and specificity (range: 0.68-0.98) ranged widely and did not correlate with years of experience (p ≥ 0.49). Algorithm performance (AUC = 0.93) also matched or exceeded the sensitivity (range: 0.78-1.00) and specificity (range: 0.32-0.87) of 6 LAC DHS optometrists in the subsets of the test dataset they graded based on expert panel reference labels. A DL algorithm for detecting referable glaucoma developed using patient-level data provided by trained LAC DHS optometrists approximates or exceeds performance by ophthalmologists and optometrists, who exhibit variable sensitivity and specificity unrelated to experience level. Implementation of this algorithm in screening workflows could help reallocate eye care resources and provide more reproducible and timely glaucoma care.

Automated peripheral nerve segmentation for MR-neurography

BackgroundMagnetic resonance neurography (MRN) is increasingly used as a diagnostic tool for peripheral neuropathies. Quantitative measures enhance MRN interpretation but require nerve segmentation which is time-consuming and error-prone and has not become clinical routine. In this study, we applied neural networks for the automated segmentation of peripheral nerves.MethodsA neural segmentation network was trained to segment the sciatic nerve and its proximal branches on the MRN scans of the right and left upper leg of 35 healthy individuals, resulting in 70 training examples, via 5-fold cross-validation (CV). The model performance was evaluated on an independent test set of one-sided MRN scans of 60 healthy individuals.ResultsMean Dice similarity coefficient (DSC) in CV was 0.892 (95% confidence interval [CI]: 0.888–0.897) with a mean Jaccard index (JI) of 0.806 (95% CI: 0.799–0.814) and mean Hausdorff distance (HD) of 2.146 (95% CI: 2.184–2.208). For the independent test set, DSC and JI were lower while HD was higher, with a mean DSC of 0.789 (95% CI: 0.760–0.815), mean JI of 0.672 (95% CI: 0.642–0.699), and mean HD of 2.118 (95% CI: 2.047–2.190).ConclusionThe deep learning-based segmentation model showed a good performance for the task of nerve segmentation. Future work will focus on extending training data and including individuals with peripheral neuropathies in training to enable advanced peripheral nerve disease characterization.Relevance statementThe results will serve as a baseline to build upon while developing an automated quantitative MRN feature analysis framework for application in routine reading of MRN examinations.Key PointsQuantitative measures enhance MRN interpretation, requiring complex and challenging nerve segmentation.We present a deep learning-based segmentation model with good performance.Our results may serve as a baseline for clinical automated quantitative MRN segmentation.Graphical

DRpred: A Novel Deep Learning-Based Predictor for Multi-Label mRNA Subcellular Localization Prediction by Incorporating Bayesian Inferred Prior Label Relationships.

The subcellular localization of messenger RNA (mRNA) not only helps us to understand the localization regulation of gene expression but also helps to understand the relationship between RNA localization pattern and human disease mechanism, which has profound biological and medical significance. Several predictors have been proposed for predicting the subcellular localization of mRNA. However, there is still considerable room for improvement in their predictive performance, especially regarding multi-label prediction. This study proposes a novel multi-label predictor, DRpred, for mRNA subcellular localization prediction. This predictor first utilizes Bayesian networks to capture the dependencies among labels. Subsequently, it combines these dependencies with features extracted from mRNA sequences using Word2vec, forming the input for the predictor. Finally, it employs a neural network combining BiLSTM and an attention mechanism to capture the internal relationships of the input features for mRNA subcellular localization. The experimental validation on an independent test set demonstrated that DRpred obtained a competitive predictive performance in multi-label prediction and outperformed state-of-the-art predictors in predicting single subcellular localizations, obtaining accuracies of 82.14%, 93.02%, 80.37%, 94.00%, 90.58%, 84.53%, 82.01%, 79.71%, and 85.67% for the chromatin, cytoplasm, cytosol, exosome, membrane, nucleolus, nucleoplasm, nucleus, and ribosome, respectively. It is anticipated to offer profound insights for biological and medical research.

NeuroPpred-SHE: An interpretable neuropeptides prediction model based on selected features from hand-crafted features and embeddings of T5 model

Neuropeptides are the most ubiquitous neurotransmitters in the immune system, regulating various biological processes. Neuropeptides play a significant role for the discovery of new drugs and targets for nervous system disorders. Traditional experimental methods for identifying neuropeptides are time-consuming and costly. Although several computational methods have been developed to predict the neuropeptides, the accuracy is still not satisfactory due to the representability of the extracted features. In this work, we propose an efficient and interpretable model, NeuroPpred-SHE, for predicting neuropeptides by selecting the optimal feature subset from both hand-crafted features and embeddings of a protein language model. Specially, we first employed a pre-trained T5 protein language model to extract embedding features and twelve other encoding methods to extract hand-crafted features from peptide sequences, respectively. Secondly, we fused both embedding features and hand-crafted features to enhance the feature representability. Thirdly, we utilized random forest (RF), Max-Relevance and Min-Redundancy (mRMR) and eXtreme Gradient Boosting (XGBoost) methods to select the optimal feature subset from the fused features. Finally, we employed five machine learning methods (GBDT, XGBoost, SVM, MLP, and LightGBM) to build the models. Our results show that the model based on GBDT achieves the best performance. Furthermore, our final model was compared with other state-of-the-art methods on an independent test set, the results indicate that our model achieves an AUROC of 97.8 % which is higher than all the other state-of-the-art predictors. Our model is available at: https://github.com/wenjean/NeuroPpred-SHE.

Preoperative Ultrasound Radomics to Predict Posthepatectomy Liver Failure in Patients With Hepatocellular Carcinoma.

Posthepatectomy liver failure (PHLF) is a major cause of postoperative mortality in hepatocellular carcinoma (HCC) patients. The study aimed to develop a method based on the two-dimensional shear wave elastography and clinical data to evaluate the risk of PHLF in HCC patients with chronic hepatitis B. This multicenter study proposed a deep learning model (PHLF-Net) incorporating dual-modal ultrasound features and clinical indicators to predict the PHLF risk. The datasets were divided into a training cohort, an internal validation cohort, an internal independent testing cohort, and three external independent testing cohorts. Based on ResNet50 pretrained on ImageNet, PHLF-Net used a progressive training strategy with images of varying granularity and incorporated conventional B-mode and elastography images and clinical indicators related to liver reserve function. In total, 532 HCC patients who underwent hepatectomy at five hospitals were enrolled. PHLF occurred in 147 patients (27.6%, 147/532). The PHLF-Net combining dual-modal ultrasound and clinical indicators demonstrated high effectiveness for predicting PHLF, with AUCs of 0.957 and 0.923 in the internal validation and testing sets, and AUCs of 0.950, 0.860, and 1.000 in the other three independent external testing sets. The performance of PHLF-Net outperformed models of single- and dual-modal US. Preoperative ultrasound imaging combining clinical indicators can effectively predict the PHLF probability in patients with HCC. In the internal and external validation sets, PHLF-Net demonstrated its usefulness in predicting PHLF.

Exploring Molecular Heteroencoders with Latent Space Arithmetic: Atomic Descriptors and Molecular Operators.

A variational heteroencoder based on recurrent neural networks, trained with SMILES linear notations of molecular structures, was used to derive the following atomic descriptors: delta latent space vectors (DLSVs) obtained from the original SMILES of the whole molecule and the SMILES of the same molecule with the target atom replaced. Different replacements were explored, namely, changing the atomic element, replacement with a character of the model vocabulary not used in the training set, or the removal of the target atom from the SMILES. Unsupervised mapping of the DLSV descriptors with t-distributed stochastic neighbor embedding (t-SNE) revealed a remarkable clustering according to the atomic element, hybridization, atomic type, and aromaticity. Atomic DLSV descriptors were used to train machine learning (ML) models to predict 19F NMR chemical shifts. An R2 of up to 0.89 and mean absolute errors of up to 5.5 ppm were obtained for an independent test set of 1046 molecules with random forests or a gradient-boosting regressor. Intermediate representations from a Transformer model yielded comparable results. Furthermore, DLSVs were applied as molecular operators in the latent space: the DLSV of a halogenation (H→F substitution) was summed to the LSVs of 4135 new molecules with no fluorine atom and decoded into SMILES, yielding 99% of valid SMILES, with 75% of the SMILES incorporating fluorine and 56% of the structures incorporating fluorine with no other structural change.

A semen-specific deoxyribonucleic acid methylation model for epigenetic age estimation and its robustness under environmental challenges.

In forensic investigations, semen samples are a common form of biological evidence, especially in cases involving sexual assault. Therefore, accurately estimating the age of an individual is crucial in criminal cases. This study presents a novel age estimation model based on semen-specific CpG methylation patterns. A multiplex panel was developed, consisting of 12 CpG sites (PARP14, C5orf25, cg23488376, MXRA5, PFKFB3, DLL1, NOX4, cg12837463, TTC7B, KCNA7, NKX2-1, and SYNE4), which exhibit strong correlations with age. Additionally, this study investigates the resilience of these methylation markers under simulated environmental challenges. We collected ejaculate samples from a diverse cohort of 115 male individuals, aged 20-71 years, who underwent deoxyribonucleic acid extraction and bisulfite conversion. Methylation levels of the selected CpG sites were assessed using a SNaPshot assay, which revealed significant correlations with chronological age. We developed and validated two robust age estimation models through stepwise and enter regression analyses, achieving reliable accuracy with mean absolute errors ranging from 3.81 to 4.1 years. Additionally, the study also investigated the robustness of semen stains under diverse environmental conditions, including fabric type, washing, hematin exposure, and UV-C light. The selected methylation markers demonstrated remarkable resilience despite the challenges posed by washing procedures and environmental exposure, confirming their potential for age estimation in forensic genetics. This research presents successful age estimation models, emphasizing the strong correlations between methylation levels and chronological age. The proposed methodology's accuracy is affirmed through model validation on an independent test set, while also highlighting the resilience of semen stains on fabrics under varying storage and washing conditions.

MycoAI: Fast and accurate taxonomic classification for fungal ITS sequences.

Efficient and accurate classification of DNA barcode data is crucial for large-scale fungal biodiversity studies. However, existing methods are either computationally expensive or lack accuracy. Previous research has demonstrated the potential of deep learning in this domain, successfully training neural networks for biological sequence classification. We introduce the MycoAI Python package, featuring various deep learning models such as BERT and CNN tailored for fungal Internal Transcribed Spacer (ITS) sequences. We explore different neural architecture designs and encoding methods to identify optimal models. By employing a multi-head output architecture and multi-level hierarchical label smoothing, MycoAI effectively generalizes across the taxonomic hierarchy. Using over 5 million labelled sequences from the UNITE database, we develop two models: MycoAI-BERT and MycoAI-CNN. While we emphasize the necessity of verifying classification results by AI models due to insufficient reference data, MycoAI still exhibits substantial potential. When benchmarked against existing classifiers such as DNABarcoder and RDP on two independent test sets with labels present in the training dataset, MycoAI models demonstrate high accuracy at the genus and higher taxonomic levels, with MycoAI-CNN being the fastest and most accurate. In terms of efficiency, MycoAI models can classify over 300,000 sequences within 5 min. We publicly release the MycoAI models, enabling mycologists to classify their ITS barcode data efficiently. Additionally, MycoAI serves as a platform for developing further deep learning-based classification methods. The source code for MycoAI is available under the MIT Licence at https://github.com/MycoAI/MycoAI.

An integrated and rapid evaluation of Curcumae Radix from different botanical origins based on chemical components, antiplatelet aggregation effect and Fourier transform near-infrared spectroscopy

Curcumae Radix (CR) is a widely used traditional Chinese medicine with significant pharmaceutical importance, including enhancing blood circulation and addressing blood stasis. This study aims to establish an integrated and rapid quality assessment method for CR from various botanical origins, based on chemical components, antiplatelet aggregation effects, and Fourier transform near-infrared (FT-NIR) spectroscopy combined with multivariate algorithms. Firstly, ultra-performance liquid chromatography-photodiode array (UPLC-PDA) combined with chemometric analyses was used to examine variations in the chemical profiles of CR. Secondly, the activation effect on blood circulation of CR was assessed using an in vitro antiplatelet aggregation assay. The studies revealed significant variations in chemical profiles and antiplatelet aggregation effects among CR samples from different botanical origins, with constituents such as germacrone, β-elemene, bisdemethoxycurcumin, demethoxycurcumin, and curcumin showing a positive correlation with antiplatelet aggregation biopotency.Thirdly, FT-NIR spectroscopy was integrated with various machine learning algorithms, including Artificial Neural Network (ANN), K-Nearest Neighbors (KNN), Logistic Regression (LR), Support Vector Machine (SVM), and Subspace K-Nearest Neighbors (Subspace KNN), to classify CR samples from four distinct sources. The result showed that FT-NIR combined with KNN and SVM classification algorithms after SNV and MSC preprocessing successfully distinguished CR samples from four plant sources with an accuracy of 100%.Finally, Quantitative models for active constituents and antiplatelet aggregation bioactivity were developed by optimizing the partial least squares (PLS) model with interval combination optimization (ICO) and competitive adaptive reweighted sampling (CARS) techniques. The CARS-PLS model achieved the best predictive performance across all five components. The coefficient of determination (R2p) and root mean square error (RMSEP) in the independent test sets were 0.9708 and 0.2098, 0.8744 and 0.2065, 0.9511 and 0.0034, 0.9803 and 0.0066, 0.9567 and 0.0172 for germacrone, β-elemene, bisdemethoxycurcumin, demethoxycurcumin and curcumin, respectively. The ICO-PLS model demonstrated superior predictive capabilities for antiplatelet aggregation biotency, achieving an R2p of 0.9010, and an RMSEP of 0.5370.This study provides a valuable reference for the quality evaluation of CR in a more rapid and comprehensive manner.

Building machine learning prediction models for well-being using predictors from the exposome and genome in a population cohort

Effective personalized well-being interventions require the ability to predict who will thrive or not, and the understanding of underlying mechanisms. Here, using longitudinal data of a large population cohort (the Netherlands Twin Register, collected 1991–2022), we aim to build machine learning prediction models for adult well-being from the exposome and genome, and identify the most predictive factors (N between 702 and 5874). The specific exposome was captured by parent and self-reports of psychosocial factors from childhood to adulthood, the genome was described by polygenic scores, and the general exposome was captured by linkage of participants’ postal codes to objective, registry-based exposures. Not the genome (R2 = −0.007 [−0.026–0.010]), but the general exposome (R2 = 0.047 [0.015–0.076]) and especially the specific exposome (R2 = 0.702 [0.637–0.753]) were predictive of well-being in an independent test set. Adding the genome (P = 0.334) and general exposome (P = 0.695) independently or jointly (P = 0.029) beyond the specific exposome did not improve prediction. Risk/protective factors such as optimism, personality, social support and neighborhood housing characteristics were most predictive. Our findings highlight the importance of longitudinal monitoring and promises of different data modalities for well-being prediction.

Effect of abdominal adipose tissue glucose uptake on brain aging.

Abdominal adipose tissue (AT) mass has adverse effects on the brain. This study aimed to investigate the effect of glucose uptake by abdominal AT on brain aging. Three-hundred twenty-five participants underwent total-body positron emission tomography scan. Brain age was estimated in an independent test set (n=98) using a support vector regression model that was built using a training set (n=227). Effects of abdominal subcutaneous and visceral AT (SAT/VAT) glucose uptake on brain age delta were evaluated using linear regression. Higher VAT glucose uptake was linked to negative brain age delta across all subgroups. Higher SAT glucose uptake was associated with negative brain age delta in lean individuals. In contrast, increased SAT glucose uptake demonstrated positive trends with brain age delta in female and overweight/obese participants. Increased glucose uptake of the abdominal VAT has positive influences on the brain, while SAT may not have such influences, except for lean individuals. Higher glucose uptake of the visceral adipose tissue was linked to decelerated brain aging. Higher glucose uptake of the subcutaneous adipose tissue (SAT) was associated with negative brain age delta in lean individuals. Faster brain aging was associated with increased glucose uptake of the SAT in female and overweight and obese individuals.

Comprehensive Analysis and In Vitro Verification of Endothelial-Mesenchymal Transition-Related Genes in Moyamoya Disease.

Moyamoya disease (MMD) is a rare, chronic, and progressive cerebrovascular disorder with unclear underlying causes and mechanisms. Previous studies suggest a potential involvement of endothelial-mesenchymal transition (EndMT) in the pathogenesis of MMD. This study aimed to explore the contribution of EndMT-related genes (ERGs) in MMD. Two datasets, GSE141022 and GSE157628, were integrated as the training set after batch effects removal. Differentially expressed ERGs were identified between MMD and control groups. Functional enrichment analysis and immune infiltration analysis were further performed. LASSO regression was used for hub MMD-related ERG selection. Consensus clustering was used for MMD subtype classification based on these hub MMD-related ERGs. Molecular characteristics between MMD subtypes were analyzed using WGCNA. PPI network was used to illuminate the genetic relationship. The hub MMD-related ERGs were validated in an independent testing set, GSE189993. The nomogram model was constructed and evaluated using ROC curves and calibration plots. Additionally, CCK-8, EdU, wound healing, and western blot were performed to confirm the function of the hub MMD-related ERGs. A total of 107 DE-ERGs were identified. Functional enrichment analysis showed these genes were associated with EndMT and immune response. The infiltrating levels of immune cells were commonly higher in the MMD group. LASSO regression identified 12 hub MMD-related ERGs, leading to the identification of two MMD subtypes. Four ERGs emerged as the final hub MMD-related ERGs after validation in the testing set, including CCL21, CEBPA, KRT18, and TNFRSF11A. The nomogram model exhibited excellent discrimination ability. In vitro experiments showed that CCL21, CEBPA, KRT18, and TNFRSF11A could promote proliferation, migration, and EndMT. This study investigated the potential role of EndMT in MMD and identified four hub MMD-related ERGs, providing potential therapeutic targets for MMD treatment.

Discovery of potential antidiabetic peptides using deep learning

Antidiabetic peptides (ADPs), peptides with potential antidiabetic activity, hold significant importance in the treatment and control of diabetes. Despite their therapeutic potential, the discovery and prediction of ADPs remain challenging due to limited data, the complex nature of peptide functions, and the expensive and time-consuming nature of traditional wet lab experiments. This study aims to address these challenges by exploring methods for the discovery and prediction of ADPs using advanced deep learning techniques. Specifically, we developed two models: a single-channel CNN and a three-channel neural network (CNN + RNN + Bi-LSTM). ADPs were primarily gathered from the BioDADPep database, alongside thousands of non-ADPs sourced from anticancer, antibacterial, and antiviral peptide datasets. Subsequently, data preprocessing was performed with the evolutionary scale model (ESM-2), followed by model training and evaluation through 10-fold cross-validation. Furthermore, this work collected a series of newly published ADPs as an independent test set through literature review, and found that the CNN model achieved the highest accuracy (90.48 %) in predicting the independent test set, surpassing existing ADP prediction tools. Finally, the application of the model was considered. SeqGAN was used to generate new candidate ADPs, followed by screening with the constructed CNN model. Selected peptides were then evaluated using physicochemical property prediction and structural forecasts for pharmaceutical potential. In summary, this study not only established robust ADP prediction models but also employed these models to screen a batch of potential ADPs, addressing a critical need in the field of peptide-based antidiabetic research.

I-Biomarker CaDx: A circulating miRNA-based multi-cancer detection test with generative and explainable AI for colorectal cancer.

64 Background: Colorectal cancer (CRC), the second most deadly cancer, underscores the critical need for early detection to significantly improve treatment outcomes and survival rates. Colonoscopy, flexible sigmoidoscopy, and the fecal immunochemical tests (FIT), often fail to capture early disease, significantly decreasing the survival chance. Circulating miRNA is a promising non-invasive biomarker for various cancers, including CRC. The i-Biomarker CaDx (patent pending) is an innovative multi-cancer early detection and diagnosis platform, covering 32 cancer types with an impressive accuracy rate of 99-100%. This study focuses on the effectiveness of i-Biomarker CaDx for CRC, employing significant data collection and leveraging insights from Generative AI (GenAI) and Explainable Artificial Intelligence (XAI). Methods: We collected various available datasets (e.g. GSE106817, etc.) encompassing 305 colorectal cancer patients and a matching number of healthy controls. Microarray was employed to profile circulating miRNAs. We used various classification paradigms, e.g., decision trees, neural networks, etc. Hyperparameter optimization was conducted, and the best-performing classifiers were amalgamated in the final model, weighted by their respective efficacies. We assessed i-Biomarker CaDx's performance through cross-validation and independent test sets. The XAI provided detailed insights into miRNA variations associated with personalized diagnostic. GenAI integrated the test into the screening and early detection best practice workflows and performs Functional Analysis. Results: i-Biomarker CaDx shows exceptional diagnostic accuracy of 99-100%, surpassing conventional colorectal cancer diagnostic methods like colonoscopy, flexible sigmoidoscopy, and FIT. It can also be used for personalized treatment response monitoring. I-Biomarker effectively deciphers intricate miRNA relationships relevant to colorectal cancer using XAI. Thus, our analyses shed light on miRNA patterns and their associations with colorectal cancer, enhancing the understanding of the underlying molecular complexities of the disease. It goes further by explaining the personalized cellular and molecular mechanisms involved, and it will be integrated into the corresponding medical workflows. Conclusions: Our AI-powered multi-cancer early diagnostic platform demonstrates outstanding performance in detecting CRC, surpassing traditional diagnostic methods like colonoscopy, flexible sigmoidoscopy, and FIT. XAI allows for in-depth exploration of miRNA alterations and their impact on CRC, enriching our interpretations at both the population and individual levels. These findings underscore the substantial potential of i-Biomarker CaDx as a transformative, non-invasive diagnostic tool for CRC.

A novel deep learning model for obstructive sleep apnea diagnosis: hybrid CNN-Transformer approach for radar-based detection of apnea-hypopnea events.

The demand for cost-effective and accessible alternatives to polysomnography (PSG), the conventional diagnostic method for obstructive sleep apnea (OSA), has surged. In this study, we have developed and validated a deep learning model for detecting apnea-hypopnea events using radar data. We conducted a single-center prospective cohort study, dividing participants with suspected sleep-disordered breathing into development and temporally independent test sets. Utilizing a hybrid CNN-Transformer architecture, we performed 5-fold cross-validation on the development set to develop and subsequently validate the model. Evaluation metrics included sensitivity for event detection, mean absolute error (MAE), intraclass correlation coefficient (ICC), and Pearson correlation coefficient (r) for apnea-hypopnea index (AHI) estimation. Linearly weighted kappa statistics (κ) assessed OSA severity. The development set comprised 54 participants (July 2021-May 2022), while the test set included 35 participants (June 2022-June 2023). In the test set, our model achieved an event detection sensitivity of 67.2% (95% CI: 65.8%, 68.5%) and demonstrated a MAE of 7.54 (95% CI: 5.36, 9.72), indicating good agreement (ICC = 0.889 [95% CI: 0.792, 0.942]) and a strong correlation (r = 0.892 [95% CI: 0.795, 0.945]) with the ground truth for AHI estimation. Furthermore, OSA severity estimation showed substantial agreement (κ = 0.780 [95% CI: 0.658, 0.903]). Our study highlights radar sensors and advanced AI models' potential to improve OSA diagnosis, paving the path for future radar-based diagnostic models in sleep medicine research.

AI-based automated segmentation for ovarian/adnexal masses and their internal components on ultrasound imaging.

Segmentation of ovarian/adnexal masses from surrounding tissue on ultrasound images is a challenging task. The separation of masses into different components may also be important for radiomic feature extraction. Our study aimed to develop an artificial intelligence-based automatic segmentation method for transvaginal ultrasound images that (1)outlines the exterior boundary of adnexal masses and (2)separates internal components. A retrospective ultrasound imaging database of adnexal masses was reviewed for exclusion criteria at the patient, mass, and image levels, with one image per mass. The resulting 54 adnexal masses (36 benign/18 malignant) from 53 patients were separated by patient into training (26 benign/12 malignant) and independent test (10 benign/6 malignant) sets. U-net segmentation performance on test images compared to expert detailed outlines was measured using the Dice similarity coefficient (DSC) and the ratio of the Hausdorff distance to the effective diameter of the outline ( ) for each mass. Subsequently, in discovery mode, a two-level fuzzy c-means (FCM) unsupervised clustering approach was used to separate the pixels within masses belonging to hypoechoic or hyperechoic components. The DSC (median [95% confidence interval]) was 0.91 [0.78, 0.96], and was 0.04 [0.01, 0.12], indicating strong agreement with expert outlines. Clinical review of the internal separation of masses into echogenic components demonstrated a strong association with mass characteristics. A combined U-net and FCM algorithm for automatic segmentation of adnexal masses and their internal components achieved excellent results compared with expert outlines and review, supporting future radiomic feature-based classification of the masses by components.