Sci Rep. 2024 Jul 11;14(1):16003. doi: 10.1038/s41598-024-66797-1.

ABSTRACT

In order to extract more important morphological features of neuron images and achieve accurate classification of the neuron type, a method is proposed that uses Sugeno fuzzy integral integration of three optimized deep learning models, namely AlexNet, VGG11_bn, and ResNet-50. Firstly, using the pre-trained model of AlexNet and the output layer is fine-tuned to improve the model's performance. Secondly, in the VGG11_bn network, Global Average Pooling (GAP) is adopted to replace the traditional fully connected layer to reduce the number of parameters. Additionally, the generalization ability of the model is improved by transfer learning. Thirdly, the SE(squeeze and excitation) module is added to the ResNet-50 variant ResNeXt-50 to adjust the channel weight and capture the key information of the input data. The GELU activation function is used to better fit the data distribution. Finally, Sugeno fuzzy integral is used to fuse the output of each model to get the final classification result. The experimental results showed that on the Img_raw, Img_resample and Img_XYalign dataset, the accuracy of 4-category classification reached 98.04%, 91.75% and 93.13%, respectively, and the accuracy of 12-category classification reached 97.82%, 85.68% and 87.60%, respectively. The proposed method has good classification performance in the morphological classification of neurons.

PMID:38992081 | DOI:10.1038/s41598-024-66797-1

Sci Rep. 2024 Jul 11;14(1):16035. doi: 10.1038/s41598-024-67107-5.

ABSTRACT

In the realm of cloud computing, ensuring the dependability and robustness of software systems is paramount. The intricate and evolving nature of cloud infrastructures, however, presents substantial obstacles in the pre-emptive identification and rectification of software anomalies. This study introduces an innovative methodology that amalgamates hybrid optimization algorithms with Neural Networks (NN) to refine the prediction of software malfunctions. The core objective is to augment the purity metric of our method across diverse operational conditions. This is accomplished through the utilization of two distinct optimization algorithms: the Yellow Saddle Goat Fish Algorithm (YSGA), which is instrumental in the discernment of pivotal features linked to software failures, and the Grasshopper Optimization Algorithm (GOA), which further polishes the feature compilation. These features are then processed by Neural Networks (NN), capitalizing on their proficiency in deciphering intricate data patterns and interconnections. The NNs are integral to the classification of instances predicated on the ascertained features. Our evaluation, conducted using the Failure-Dataset-OpenStack database and MATLAB Software, demonstrates that the hybrid optimization strategy employed for feature selection significantly curtails complexity and expedites processing.

PMID:38992079 | DOI:10.1038/s41598-024-67107-5

Sci Rep. 2024 Jul 11;14(1):16038. doi: 10.1038/s41598-024-66182-y.

ABSTRACT

In transportation, roads sometimes have cracks due to overloading and other reasons, which seriously affect driving safety, and it is crucial to identify and fill road cracks in time. Aiming at the defects of existing semantic segmentation models that have degraded the segmentation performance of road crack images and the standard convolution makes it challenging to capture the spatial and channel coupling relationship between pixels. It is difficult to differentiate crack pixels from background pixels in complex backgrounds; this paper proposes a semantic segmentation model for road cracks that combines channel-spatial convolution with the aggregation of frequency features. A new convolutional block is proposed to accurately identify cracked pixels by grouping spatial displacements and convolutional kernel weight dynamization while modeling pixel spatial relationships linked to channel features. To enhance the contrast of crack edges, a frequency domain feature aggregation module is proposed, which uses a simple windowing strategy to solve the problem of mismatch of frequency domain inputs and, at the same time, takes into account the effect of the frequency imaginary part on the features to model the deep frequency features effectively. Finally, a feature refinement module is designed to refine the semantic features to improve the segmentation accuracy. Many experiments have proved that the model proposed in this paper has better performance and more application potential than the current popular general model.

PMID:38992078 | DOI:10.1038/s41598-024-66182-y

AJNR Am J Neuroradiol. 2024 Jul 11. doi: 10.3174/ajnr.A8335. Online ahead of print.

ABSTRACT

BACKGROUND AND PURPOSE: CT imaging exposes patients to ionizing radiation. MR imaging is radiation free but previously has not been able to produce diagnostic-quality images of bone on a timeline suitable for clinical use. We developed automated motion correction and use deep learning to generate pseudo-CT images from MR images. We aim to evaluate whether motion-corrected pseudo-CT produces cranial images that have potential to be acceptable for clinical use.

MATERIALS AND METHODS: Patients younger than age 18 who underwent CT imaging of the head for either trauma or evaluation of cranial suture patency were recruited. Subjects underwent a 5-minute golden-angle stack-of-stars radial volumetric interpolated breath-hold MR image. Motion correction was applied to the MR imaging followed by a deep learning-based method to generate pseudo-CT images. CT and pseudo-CT images were evaluated and, based on indication for imaging, either presence of skull fracture or cranial suture patency was first recorded while viewing the MR imaging-based pseudo-CT and then recorded while viewing the clinical CT.

RESULTS: A total of 12 patients underwent CT and MR imaging to evaluate suture patency, and 60 patients underwent CT and MR imaging for evaluation of head trauma. For cranial suture patency, pseudo-CT had 100% specificity and 100% sensitivity for the identification of suture closure. For identification of skull fractures, pseudo-CT had 100% specificity and 90% sensitivity.

CONCLUSIONS: Our early results show that automated motion-corrected and deep learning-generated pseudo-CT images of the pediatric skull have potential for clinical use and offer a high level of diagnostic accuracy when compared with standard CT scans.

PMID:38991771 | DOI:10.3174/ajnr.A8335

Neuro Oncol. 2024 Jul 11:noae113. doi: 10.1093/neuonc/noae113. Online ahead of print.

ABSTRACT

BACKGROUND: Artificial intelligence has been proposed for brain metastasis (BM) segmentation but it has not been fully clinically validated. The aim of this study was to develop and evaluate a system for BM segmentation.

METHODS: A deep-learning-based BM segmentation system (BMSS) was developed using contrast-enhanced MR images from 488 patients with 10,338 brain metastases. A randomized crossover, multi-reader study was then conducted to evaluate the performance of the BMSS for BM segmentation using data prospectively collected from 50 patients with 203 metastases at five centers. Five radiology residents and five attending radiologists were randomly assigned to contour the same prospective set in assisted and unassisted modes. Aided and unaided Dice similarity coefficients (DSCs) and contouring times per lesion were compared.

RESULTS: The BMSS alone yielded a median DSC of 0.91 (95% confidence interval, 0.90-0.92) in the multi-center set and showed comparable performance between the internal and external sets (p = 0.67). With BMSS assistance, the readers increased the median DSC from 0.87 (0.87-0.88) to 0.92 (0.92-0.92) (p < 0.001) with a median time saving of 42% (40-45%) per lesion. Resident readers showed a greater improvement than attending readers in contouring accuracy (improved median DSC, 0.05 [0.05-0.05] vs. 0.03 [0.03-0.03]; p < 0.001), but a similar time reduction (reduced median time, 44% [40-47%] vs. 40% [37-44%]; p = 0.92) with BMSS assistance.

CONCLUSIONS: The BMSS can be optimally applied to improve the efficiency of brain metastasis delineation in clinical practice.

PMID:38991556 | DOI:10.1093/neuonc/noae113

Artif Intell Med. 2024 Jun 28;154:102927. doi: 10.1016/j.artmed.2024.102927. Online ahead of print.

ABSTRACT

Stroke stands as a major global health issue, causing high death and disability rates and significant social and economic burdens. The effectiveness of existing stroke risk assessment methods is questionable due to their use of inconsistent and varying biomarkers, which may lead to unpredictable risk evaluations. This study introduces an automatic deep learning-based system for predicting stroke risk (both ischemic and hemorrhagic) and estimating the time frame of its occurrence, utilizing a comprehensive set of known retinal biomarkers from fundus images. Our system, tested on the UK Biobank and DRSSW datasets, achieved AUROC scores of 0.83 (95% CI: 0.79-0.85) and 0.93 (95% CI: 0.9-0.95), respectively. These results not only highlight our system's advantage over established benchmarks but also underscore the predictive power of retinal biomarkers in assessing stroke risk and the unique effectiveness of each biomarker. Additionally, the correlation between retinal biomarkers and cardiovascular diseases broadens the potential application of our system, making it a versatile tool for predicting a wide range of cardiovascular conditions.

PMID:38991398 | DOI:10.1016/j.artmed.2024.102927

J Chem Inf Model. 2024 Jul 11. doi: 10.1021/acs.jcim.4c00783. Online ahead of print.

ABSTRACT

Long-range allosteric communication between distant sites and active sites in proteins is central to biological regulation but still poorly characterized, limiting the development of protein engineering and drug design. Addressing this gap, NRIMD is an open-access web server for analyzing long-range interactions in proteins from molecular dynamics (MD) simulations, such as the effect of mutations at distal sites or allosteric ligand binding at allosteric sites on the active center. Based on our recent works on neural relational inference using graph neural networks, this cloud-based web server accepts MD simulation data on any length of residues in the alpha-carbon skeleton format from mainstream MD software. The input trajectory data are validated at the frontend deployed on the cloud and then processed on the backend deployed on a high-performance computer system with a collection of complementary tools. The web server provides a one-stop-shop MD analysis platform to predict long-range interactions and their paths between distant sites and active sites. It provides a user-friendly interface for detailed analysis and visualization. To the best of our knowledge, NRIMD is the first-of-its-kind online service to provide comprehensive long-range interaction analysis on MD simulations, which significantly lowers the barrier of predictions on protein long-range interactions using deep learning. The NRIMD web server is publicly available at https://nrimd.luddy.indianapolis.iu.edu/.

PMID:38991149 | DOI:10.1021/acs.jcim.4c00783

J Chem Theory Comput. 2024 Jul 11. doi: 10.1021/acs.jctc.4c00503. Online ahead of print.

ABSTRACT

Understanding drug residence times in target proteins is key to improving drug efficacy and understanding target recognition in biochemistry. While drug residence time is just as important as binding affinity, atomic-level understanding of drug residence times through molecular dynamics (MD) simulations has been difficult primarily due to the extremely long time scales. Recent advances in rare event sampling have allowed us to reach these time scales, yet predicting protein-ligand residence times remains a significant challenge. Here we present a semi-automated protocol to calculate the ligand residence times across 12 orders of magnitude of time scales. In our proposed framework, we integrate a deep learning-based method, the state predictive information bottleneck (SPIB), to learn an approximate reaction coordinate (RC) and use it to guide the enhanced sampling method metadynamics. We demonstrate the performance of our algorithm by applying it to six different protein-ligand complexes with available benchmark residence times, including the dissociation of the widely studied anticancer drug Imatinib (Gleevec) from both wild-type Abl kinase and drug-resistant mutants. We show how our protocol can recover quantitatively accurate residence times, potentially opening avenues for deeper insights into drug development possibilities and ligand recognition mechanisms.

PMID:38991145 | DOI:10.1021/acs.jctc.4c00503

JBJS Rev. 2024 Jul 11;12(7). doi: 10.2106/JBJS.RVW.24.00057. eCollection 2024 Jul 1.

ABSTRACT

» Artificial intelligence is an umbrella term for computational calculations that are designed to mimic human intelligence and problem-solving capabilities, although in the future, this may become an incomplete definition. Machine learning (ML) encompasses the development of algorithms or predictive models that generate outputs without explicit instructions, assisting in clinical predictions based on large data sets. Deep learning is a subset of ML that utilizes layers of networks that use various inter-relational connections to define and generalize data.» ML algorithms can enhance radiomics techniques for improved image evaluation and diagnosis. While ML shows promise with the advent of radiomics, there are still obstacles to overcome.» Several calculators leveraging ML algorithms have been developed to predict survival in primary sarcomas and metastatic bone disease utilizing patient-specific data. While these models often report exceptionally accurate performance, it is crucial to evaluate their robustness using standardized guidelines.» While increased computing power suggests continuous improvement of ML algorithms, these advancements must be balanced against challenges such as diversifying data, addressing ethical concerns, and enhancing model interpretability.

PMID:38991098 | DOI:10.2106/JBJS.RVW.24.00057

PLoS Comput Biol. 2024 Jul 11;20(7):e1011953. doi: 10.1371/journal.pcbi.1011953. Online ahead of print.

ABSTRACT

With recent methodological advances in the field of computational protein design, in particular those based on deep learning, there is an increasing need for frameworks that allow for coherent, direct integration of different models and objective functions into the generative design process. Here we demonstrate how evolutionary multiobjective optimization techniques can be adapted to provide such an approach. With the established Non-dominated Sorting Genetic Algorithm II (NSGA-II) as the optimization framework, we use AlphaFold2 and ProteinMPNN confidence metrics to define the objective space, and a mutation operator composed of ESM-1v and ProteinMPNN to rank and then redesign the least favorable positions. Using the two-state design problem of the foldswitching protein RfaH as an in-depth case study, and PapD and calmodulin as examples of higher-dimensional design problems, we show that the evolutionary multiobjective optimization approach leads to significant reduction in the bias and variance in RfaH native sequence recovery, compared to a direct application of ProteinMPNN. We suggest that this improvement is due to three factors: (i) the use of an informative mutation operator that accelerates the sequence space exploration, (ii) the parallel, iterative design process inherent to the genetic algorithm that improves upon the ProteinMPNN autoregressive sequence decoding scheme, and (iii) the explicit approximation of the Pareto front that leads to optimal design candidates representing diverse tradeoff conditions. We anticipate this approach to be readily adaptable to different models and broadly relevant for protein design tasks with complex specifications.

PMID:38991035 | DOI:10.1371/journal.pcbi.1011953

PLoS One. 2024 Jul 11;19(7):e0303462. doi: 10.1371/journal.pone.0303462. eCollection 2024.

ABSTRACT

Nowadays, federated learning is one of the most prominent choices for making decisions. A significant benefit of federated learning is that, unlike deep learning, it is not necessary to share data samples with the model owner. The weight of the global model in traditional federated learning is created by averaging the weights of all clients or sites. In the proposed work, a novel method has been discussed to generate an optimized base model without hampering its performance, which is based on a genetic algorithm. Chromosome representation, crossover, and mutation-all the intermediate operations of the genetic algorithm have been illustrated with useful examples. After applying the genetic algorithm, there is a significant improvement in inference time and a huge reduction in storage space. Therefore, the model can be easily deployed on resource-constrained devices. For the experimental work, sports data has been used in balanced and unbalanced scenarios with various numbers of clients in a federated learning environment. In addition, we have used four famous deep learning architectures, such as AlexNet, VGG19, ResNet50, and EfficientNetB3, as the base model. We have achieved 92.34% accuracy with 9 clients in the balanced data set by using EfficientNetB3 as the base model using a GA-based approach. Moreover, after applying the genetic algorithm to optimize EfficientNetB3, there is an improvement in inference time and storage space by 20% and 2.35%, respectively.

PMID:38990969 | DOI:10.1371/journal.pone.0303462

PLoS One. 2024 Jul 11;19(7):e0301908. doi: 10.1371/journal.pone.0301908. eCollection 2024.

ABSTRACT

Real-time security surveillance and identity matching using face detection and recognition are central research areas within computer vision. The classical facial detection techniques include Haar-like, MTCNN, AdaBoost, and others. These techniques employ template matching and geometric facial features for detecting faces, striving for a balance between detection time and accuracy. To address this issue, the current research presents an enhanced FaceNet network. The RetinaFace is employed to perform expeditious face detection and alignment. Subsequently, FaceNet, with an improved loss function is used to achieve face verification and recognition with high accuracy. The presented work involves a comparative evaluation of the proposed network framework against both traditional and deep learning techniques in terms of face detection and recognition performance. The experimental findings demonstrate that an enhanced FaceNet can successfully meet the real-time facial recognition requirements, and the accuracy of face recognition is 99.86% which fulfills the actual requirement. Consequently, the proposed solution holds significant potential for applications in face detection and recognition within the education sector for real-time security surveillance.

PMID:38990958 | DOI:10.1371/journal.pone.0301908

PLoS One. 2024 Jul 11;19(7):e0304757. doi: 10.1371/journal.pone.0304757. eCollection 2024.

ABSTRACT

Recent advancements in AI, driven by big data technologies, have reshaped various industries, with a strong focus on data-driven approaches. This has resulted in remarkable progress in fields like computer vision, e-commerce, cybersecurity, and healthcare, primarily fueled by the integration of machine learning and deep learning models. Notably, the intersection of oncology and computer science has given rise to Computer-Aided Diagnosis (CAD) systems, offering vital tools to aid medical professionals in tumor detection, classification, recurrence tracking, and prognosis prediction. Breast cancer, a significant global health concern, is particularly prevalent in Asia due to diverse factors like lifestyle, genetics, environmental exposures, and healthcare accessibility. Early detection through mammography screening is critical, but the accuracy of mammograms can vary due to factors like breast composition and tumor characteristics, leading to potential misdiagnoses. To address this, an innovative CAD system leveraging deep learning and computer vision techniques was introduced. This system enhances breast cancer diagnosis by independently identifying and categorizing breast lesions, segmenting mass lesions, and classifying them based on pathology. Thorough validation using the Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM) demonstrated the CAD system's exceptional performance, with a 99% success rate in detecting and classifying breast masses. While the accuracy of detection is 98.5%, when segmenting breast masses into separate groups for examination, the method's performance was approximately 95.39%. Upon completing all the analysis, the system's classification phase yielded an overall accuracy of 99.16% for classification. The potential for this integrated framework to outperform current deep learning techniques is proposed, despite potential challenges related to the high number of trainable parameters. Ultimately, this recommended framework offers valuable support to researchers and physicians in breast cancer diagnosis by harnessing cutting-edge AI and image processing technologies, extending recent advances in deep learning to the medical domain.

PMID:38990817 | DOI:10.1371/journal.pone.0304757

IEEE J Biomed Health Inform. 2024 Jul 11;PP. doi: 10.1109/JBHI.2024.3426939. Online ahead of print.

ABSTRACT

Sleep staging is essential for sleep assessment and plays an important role in disease diagnosis, which refers to the classification of sleep epochs into different sleep stages. Polysomnography (PSG), consisting of many different physiological signals, e.g. electroencephalogram (EEG) and electrooculogram (EOG), is a gold standard for sleep staging. Although existing studies have achieved high performance on automatic sleep staging from PSG, there are still some limitations: 1) they focus on local features but ignore global features within each sleep epoch, and 2) they ignore cross-modality context relationship between EEG and EOG. In this paper, we propose CareSleepNet, a novel hybrid deep learning network for automatic sleep staging from PSG recordings. Specifically, we first design a multi-scale Convolutional-Transformer Epoch Encoder to encode both local salient wave features and global features within each sleep epoch. Then, we devise a Cross-Modality Context Encoder based on co-attention mechanism to model cross-modality context relationship between different modalities. Next, we use a Transformer-based Sequence Encoder to capture the sequential relationship among sleep epochs. Finally, the learned feature representations are fed into an epoch-level classifier to determine the sleep stages. We collected a private sleep dataset, SSND, and use two public datasets, Sleep-EDF-153 and ISRUC to evaluate the performance of CareSleepNet. The experiment results show that our CareSleepNet achieves the state-of-the-art performance on the three datasets. Moreover, we conduct ablation studies and attention visualizations to prove the effectiveness of each module and to analyze the influence of each modality.

PMID:38990749 | DOI:10.1109/JBHI.2024.3426939

IEEE/ACM Trans Comput Biol Bioinform. 2024 Jul 11;PP. doi: 10.1109/TCBB.2024.3426491. Online ahead of print.

ABSTRACT

Deep learning approaches, such as convolution neural networks (CNNs) and deep recurrent neural networks (RNNs), have been the backbone for predicting protein function, with promising state-of-the-art (SOTA) results. RNNs with an in-built ability (i) focus on past information, (ii) collect both short-and-long range dependency information, and (iii) bi-directional processing offers a strong sequential processing mechanism. CNNs, however, are confined to focusing on short-term information from both the past and the future, although they offer parallelism. Therefore, a novel bi-directional CNN that strictly complies with the sequential processing mechanism of RNNs is introduced and is used for developing a protein function prediction framework, Bi-SeqCNN. This is a sub-sequence-based framework. Further, Bi-SeqCNN + is an ensemble approach to better the prediction results. To our knowledge, this is the first time bi-directional CNNs are employed for general temporal data analysis and not just for protein sequences. The proposed architecture produces improvements up to +5.5% over contemporary SOTA methods on three benchmark protein sequence datasets. Moreover, it is substantially lighter and attain these results with (0.50-0.70 times) fewer parameters than the SOTA methods.

PMID:38990747 | DOI:10.1109/TCBB.2024.3426491

JCI Insight. 2024 Jul 11:e178578. doi: 10.1172/jci.insight.178578. Online ahead of print.

ABSTRACT

Clarifying multifactorial musculoskeletal disorder etiologies supports risk analysis and development of targeted prevention and treatment modalities. Deep learning enables comprehensive risk factor identification through systematic analysis of disease datasets but does not provide sufficient context for mechanistic understanding, limiting clinical applicability for etiological investigations. Conversely, multiscale biomechanical modeling can evaluate mechanistic etiology within the relevant biomechanical and physiological context. We propose a hybrid approach combining 3D explainable deep learning and multiscale biomechanical modeling; we applied this approach to investigate temporomandibular joint (TMJ) disorder etiology by systematically identifying risk factors and elucidating mechanistic relationships between risk factors and TMJ biomechanics and mechanobiology. Our 3D convolutional neural network recognized TMJ disorder patients through subject-specific morphological features in condylar, ramus, and chin. Driven by deep learning model outputs, biomechanical modeling revealed that small mandibular size and flat condylar shape were associated with increased TMJ disorder risk through increased joint force, decreased tissue nutrient availability and cell ATP production, and increased TMJ disc strain energy density. Combining explainable deep learning and multiscale biomechanical modeling addresses the "mechanism unknown" limitation undermining translational confidence in clinical applications of deep learning and increases methodological accessibility for smaller clinical datasets by providing the crucial biomechanical context.

PMID:38990647 | DOI:10.1172/jci.insight.178578

JMIR Form Res. 2024 Jul 11;8:e51327. doi: 10.2196/51327.

ABSTRACT

BACKGROUND: On June 23, 2022, the US Food and Drug Administration announced a JUUL ban policy, to ban all vaping and electronic cigarette products sold by Juul Labs.

OBJECTIVE: This study aims to understand public perceptions and discussions of this policy using Twitter (subsequently rebranded as X) data.

METHODS: Using the Twitter streaming application programming interface, 17,007 tweets potentially related to the JUUL ban policy were collected between June 22, 2022, and July 25, 2022. Based on 2600 hand-coded tweets, a deep learning model (RoBERTa) was trained to classify all tweets into propolicy, antipolicy, neutral, and irrelevant categories. A deep learning model (M3 model) was used to estimate basic demographics (such as age and gender) of Twitter users. Furthermore, major topics were identified using latent Dirichlet allocation modeling. A logistic regression model was used to examine the association of different Twitter users with their attitudes toward the policy.

RESULTS: Among 10,480 tweets related to the JUUL ban policy, there were similar proportions of propolicy and antipolicy tweets (n=2777, 26.5% vs n=2666, 25.44%). Major propolicy topics included "JUUL causes youth addition," "market surge of JUUL," and "health effects of JUUL." In contrast, major antipolicy topics included "cigarette should be banned instead of JUUL," "against the irrational policy," and "emotional catharsis." Twitter users older than 29 years were more likely to be propolicy (have a positive attitude toward the JUUL ban policy) than those younger than 29 years.

CONCLUSIONS: Our study showed that the public showed different responses to the JUUL ban policy, which varies depending on the demographic characteristics of Twitter users. Our findings could provide valuable information to the Food and Drug Administration for future electronic cigarette and other tobacco product regulations.

PMID:38990633 | DOI:10.2196/51327

Anal Chem. 2024 Jul 11. doi: 10.1021/acs.analchem.4c00563. Online ahead of print.

ABSTRACT

J-Resolved (J-Res) nuclear magnetic resonance (NMR) spectroscopy is pivotal in NMR-based metabolomics, but practitioners face a choice between time-consuming high-resolution (HR) experiments or shorter low-resolution (LR) experiments which exhibit significant peak overlap. Deep learning neural networks have been successfully used in many fields to enhance quality of natural images, especially with regard to resolution, and therefore offer the prospect of improving two-dimensional (2D) NMR data. Here, we introduce the J-RESRGAN, an adapted and modified generative adversarial network (GAN) for image super-resolution (SR), which we trained specifically for metabolomic J-Res spectra to enhance peak resolution. A novel symmetric loss function was introduced, exploiting the inherent vertical symmetry of J-Res NMR spectra. Model training used simulated high-resolution J-Res spectra of complex mixtures, with corresponding low-resolution spectra generated via blurring and down-sampling. Evaluation of peak pair resolvability on J-RESRGAN demonstrated remarkable improvement in resolution across a variety of samples. In simulated plasma data, 100% of peak pairs exhibited enhanced resolution in super-resolution spectra compared to their low-resolution counterparts. Similarly, enhanced resolution was observed in 80.8-100% of peak pairs in experimental plasma, 85.0-96.7% in urine, 94.4-98.9% in full fat milk, and 82.6-91.7% in orange juice. J-RESRGAN is not sample type, spectrometer or field strength dependent and improvements on previously acquired data can be seen in seconds on a standard desktop computer. We believe this demonstrates the promise of deep learning methods to enhance NMR metabolomic data, and in particular, the power of J-RESRGAN to elucidate overlapping peaks, advancing precision in a wide variety of NMR-based metabolomics studies. The model, J-RESRGAN, is openly accessible for download on GitHub at https://github.com/yanyan5420/J-RESRGAN.

PMID:38990576 | DOI:10.1021/acs.analchem.4c00563

Brief Bioinform. 2024 May 23;25(4):bbae340. doi: 10.1093/bib/bbae340.

ABSTRACT

Accurate prediction of molecular properties is fundamental in drug discovery and development, providing crucial guidance for effective drug design. A critical factor in achieving accurate molecular property prediction lies in the appropriate representation of molecular structures. Presently, prevalent deep learning-based molecular representations rely on 2D structure information as the primary molecular representation, often overlooking essential three-dimensional (3D) conformational information due to the inherent limitations of 2D structures in conveying atomic spatial relationships. In this study, we propose employing the Gram matrix as a condensed representation of 3D molecular structures and for efficient pretraining objectives. Subsequently, we leverage this matrix to construct a novel molecular representation model, Pre-GTM, which inherently encapsulates 3D information. The model accurately predicts the 3D structure of a molecule by estimating the Gram matrix. Our findings demonstrate that Pre-GTM model outperforms the baseline Graphormer model and other pretrained models in the QM9 and MoleculeNet quantitative property prediction task. The integration of the Gram matrix as a condensed representation of 3D molecular structure, incorporated into the Pre-GTM model, opens up promising avenues for its potential application across various domains of molecular research, including drug design, materials science, and chemical engineering.

PMID:38990515 | DOI:10.1093/bib/bbae340

Brief Bioinform. 2024 May 23;25(4):bbae297. doi: 10.1093/bib/bbae297.

ABSTRACT

Protein-peptide interactions (PPepIs) are vital to understanding cellular functions, which can facilitate the design of novel drugs. As an essential component in forming a PPepI, protein-peptide binding sites are the basis for understanding the mechanisms involved in PPepIs. Therefore, accurately identifying protein-peptide binding sites becomes a critical task. The traditional experimental methods for researching these binding sites are labor-intensive and time-consuming, and some computational tools have been invented to supplement it. However, these computational tools have limitations in generality or accuracy due to the need for ligand information, complex feature construction, or their reliance on modeling based on amino acid residues. To deal with the drawbacks of these computational algorithms, we describe a geometric attention-based network for peptide binding site identification (GAPS) in this work. The proposed model utilizes geometric feature engineering to construct atom representations and incorporates multiple attention mechanisms to update relevant biological features. In addition, the transfer learning strategy is implemented for leveraging the protein-protein binding sites information to enhance the protein-peptide binding sites recognition capability, taking into account the common structure and biological bias between proteins and peptides. Consequently, GAPS demonstrates the state-of-the-art performance and excellent robustness in this task. Moreover, our model exhibits exceptional performance across several extended experiments including predicting the apo protein-peptide, protein-cyclic peptide and the AlphaFold-predicted protein-peptide binding sites. These results confirm that the GAPS model is a powerful, versatile, stable method suitable for diverse binding site predictions.

PMID:38990514 | DOI:10.1093/bib/bbae297