Environ Sci Pollut Res Int. 2024 Jan 27. doi: 10.1007/s11356-024-32023-8. Online ahead of print.

ABSTRACT

Smart, secure, and environmentally friendly smart cities are all the rage in urban planning. Several technologies, including the Internet of Things (IoT) and edge computing, are used to develop smart cities. Early and accurate fire detection in a Smart city is always desirable and motivates the research community to create a more efficient model. Deep learning models are widely used for fire detection in existing research, but they encounter several issues in typical climate environments, such as foggy and normal. The proposed model lends itself to IoT applications for authentic fire surveillance because of its minimal configuration load. A hybrid Local Binary Pattern Convolutional Neural Network (LBP-CNN) and YOLO-V5 model-based fire detection model for smart cities in the foggy scenario is presented in this research. Additionally, we recommend a two-part technique for extracting features to be applied to YOLO throughout this article. Using a transfer learning technique, the first portion of the proposed approach for extracting features retrieves standard features. The section part is for retrieval of additional valuable information related to the current activity using the LBP (Local Binary Pattern) protective layer and classifications layers. This research utilizes an online Kaggle fire and smoke dataset with 13950 normal and foggy images. The proposed hybrid model is premised on a two-cascaded YOLO model. In the initial cascade, smoke and fire are detected in the normal surrounding region, and the second cascade fire is detected with density in a foggy environment. In experimental analysis, the proposed model achieved a fire and smoke detection precision rate of 96.25% for a normal setting, 93.2% for a foggy environment, and a combined detection average precision rate of 94.59%. The proposed hybrid system outperformed existing models in terms of better precision and density detection for fire and smoke.

PMID:38278999 | DOI:10.1007/s11356-024-32023-8

Sci Rep. 2024 Jan 26;14(1):2242. doi: 10.1038/s41598-023-50309-8.

ABSTRACT

Intrinsic properties of metal-organic frameworks (MOFs), such as their ultra porosity and high surface area, deem them promising solutions for problems involving gas adsorption. Nevertheless, due to their combinatorial nature, a huge number of structures is feasible which renders cumbersome the selection of the best candidates with traditional techniques. Recently, machine learning approaches have emerged as efficient tools to deal with this challenge, by allowing researchers to rapidly screen large databases of MOFs via predictive models. The performance of the latter is tightly tied to the mathematical representation of a material, thus necessitating the use of informative descriptors. In this work, a generalized framework to predict gaseous adsorption properties is presented, using as one and only descriptor the capstone of chemical information: the potential energy surface (PES). In order to be machine understandable, the PES is voxelized and subsequently a 3D convolutional neural network (CNN) is exploited to process this 3D energy image. As a proof of concept, the proposed pipeline is applied on predicting [Formula: see text] uptake in MOFs. The resulting model outperforms a conventional model built with geometric descriptors and requires two orders of magnitude less training data to reach a given level of performance. Moreover, the transferability of the approach to different host-guest systems is demonstrated, examining [Formula: see text] uptake in COFs. The generic character of the proposed methodology, inherited from the PES, renders it applicable to fields other than reticular chemistry.

PMID:38278851 | DOI:10.1038/s41598-023-50309-8

Nat Commun. 2024 Jan 26;15(1):779. doi: 10.1038/s41467-023-44503-5.

ABSTRACT

Neuronal activity-dependent transcription directs molecular processes that regulate synaptic plasticity, brain circuit development, behavioral adaptation, and long-term memory. Single cell RNA-sequencing technologies (scRNAseq) are rapidly developing and allow for the interrogation of activity-dependent transcription at cellular resolution. Here, we present NEUROeSTIMator, a deep learning model that integrates transcriptomic signals to estimate neuronal activation in a way that we demonstrate is associated with Patch-seq electrophysiological features and that is robust against differences in species, cell type, and brain region. We demonstrate this method's ability to accurately detect neuronal activity in previously published studies of single cell activity-induced gene expression. Further, we applied our model in a spatial transcriptomic study to identify unique patterns of learning-induced activity across different brain regions in male mice. Altogether, our findings establish NEUROeSTIMator as a powerful and broadly applicable tool for measuring neuronal activation, whether as a critical covariate or a primary readout of interest.

PMID:38278804 | DOI:10.1038/s41467-023-44503-5

Talanta. 2024 Jan 9;271:125650. doi: 10.1016/j.talanta.2024.125650. Online ahead of print.

ABSTRACT

Surface-enhanced Raman spectroscopy (SERS) can quickly identify molecular fingerprints and has been widely used in the field of rapid detection. However, the non-uniformity inherent in SERS substrate signals, coupled with the finite nature of the detection object, significantly hampers the advancement of SERS. Nowadays, the existing mature immunochromatographic assay (ICA) method is usually combined with SERS technology to address the defects of SERS detection. Nevertheless, the porous structure of the strip will also affect the signal uniformity during detection. Obviously, a method using SERS-ICA is needed to effectively solve signal fluctuations, improve detection accuracy, and has certain versatility. This paper introduces an internal standard method combining deep learning to predict and process Raman data. Based on the signal fluctuation of single-antigen SERS-ICA test strip, the double-antigen SERS-ICA test strip was constructed. The full spectrum Raman data of double-antigen SERS-ICA test strip was normalized by the sum of two characteristic peaks of internal standard molecules, and then processed by deep learning algorithm. The Relative Standard Deviation (RSD) of Raman data of bisphenol A was compared before and after internal standard normalization of double-antigen SERS-ICA test strip. The RSD processed by this method was increased by 3.8 times. After normalization, the prediction accuracy of Root Mean Square Error (RMSE) is improved by 2.66 times, and the prediction accuracy of R-square (R2) is increased from 0.961 to 0.994. The results showed that RMSE and R2 were used to comprehensively predict the collected data of double-antigen SERS-ICA test strip, which could effectively improve the prediction accuracy. The internal standard algorithm can effectively solve the challenges of uneven hot spots and poor signal reproducibility on the test strip to a certain extent, so as to improve the semi-quantitative accuracy.

PMID:38277967 | DOI:10.1016/j.talanta.2024.125650

Food Chem. 2024 Jan 24;443:138513. doi: 10.1016/j.foodchem.2024.138513. Online ahead of print.

ABSTRACT

Quantitative analysis of the quality constituents of Lonicera japonica (Jinyinhua [JYH]) using a feasible method provides important information on its evaluation and applications. Limitations of sample pretreatment, experimental site, and analysis time should be considered when identifying new methods. In response to these considerations, Raman spectroscopy combined with deep learning was used to establish a quantitative analysis model to determine the quality of JYH. Chlorogenic acid and total flavonoids were identified as analysis targets via network pharmacology. High performance liquid chromatograph and ultraviolet spectroscopy were used to construct standard curves for quantitative analysis. Raman spectra of JYH extracts (1200) were collected. Subsequently, models were built using partial least squares regression, Support Vector Machine, Back Propagation Neural Network, and One-dimensional Convolutional Neural Network (1D-CNN). Among these, the 1D-CNN model showed superior prediction capability and had higher accuracy (R2 = 0.971), and lower root mean square error, indicating its suitability for rapid quantitative analysis.

PMID:38277933 | DOI:10.1016/j.foodchem.2024.138513

J Alzheimers Dis. 2024 Jan 18. doi: 10.3233/JAD-230733. Online ahead of print.

ABSTRACT

BACKGROUND: Alzheimer's disease (AD), a major dementia cause, lacks effective treatment. MRI-based hippocampal volume measurement using artificial intelligence offers new insights into early diagnosis and intervention in AD progression.

OBJECTIVE: This study, involving 483 AD patients, 756 patients with mild cognitive impairment (MCI), and 968 normal controls (NC), investigated the predictive capability of MRI-based hippocampus volume measurements for AD risk using artificial intelligence and evidence-based medicine.

METHODS: Utilizing data from ADNI and OASIS-brains databases, three convolutional neural networks (InceptionResNetv2, Densenet169, and SEResNet50) were employed for automated AD classification based on structural MRI imaging. A multitask deep learning model and a densely connected 3D convolutional network were utilized. Additionally, a systematic meta-analysis explored the value of MRI-based hippocampal volume measurement in predicting AD occurrence and progression, drawing on 23 eligible articles from PubMed and Embase databases.

RESULTS: InceptionResNetv2 outperformed other networks, achieving 99.75% accuracy and 100% AUC for AD-NC classification and 99.16% accuracy and 100% AUC for MCI-NC classification. Notably, at a 512×512 size, InceptionResNetv2 demonstrated a classification accuracy of 94.29% and an AUC of 98% for AD-NC and 97.31% accuracy and 98% AUC for MCI-NC.

CONCLUSIONS: The study concludes that MRI-based hippocampal volume changes effectively predict AD onset and progression, facilitating early intervention and prevention.

PMID:38277290 | DOI:10.3233/JAD-230733

IEEE Trans Pattern Anal Mach Intell. 2024 Jan 26;PP. doi: 10.1109/TPAMI.2024.3359087. Online ahead of print.

ABSTRACT

Image reconstruction from incomplete measurements is one basic task in imaging. While supervised deep learning has emerged as a powerful tool for image reconstruction in recent years, its applicability is limited by its prerequisite on a large number of latent images for model training. To extend the application of deep learning to the imaging tasks where acquisition of latent images is challenging, this paper proposes an unsupervised deep learning method that trains a deep model for image reconstruction with the access limited to measurement data. We develop a Siamese network whose twin sub-networks perform reconstruction cooperatively on a pair of complementary spaces: the null space of the measurement matrix and the range space of its pseudo inverse. The Siamese network is trained by a self-supervised loss with three terms: a data consistency loss over available measurements in the range space, a data consistency loss between intermediate results in the null space, and a mutual consistency loss on the predictions of the twin sub-networks in the full space. The proposed method is applied to four imaging tasks from different applications, and extensive experiments have shown its advantages over existing unsupervised solutions.

PMID:38277253 | DOI:10.1109/TPAMI.2024.3359087

IEEE Trans Med Imaging. 2024 Jan 26;PP. doi: 10.1109/TMI.2024.3358955. Online ahead of print.

ABSTRACT

Deep learning models often need sufficient supervision (i.e. labelled data) in order to be trained effectively. By contrast, humans can swiftly learn to identify important anatomy in medical images like MRI and CT scans, with minimal guidance. This recognition capability easily generalises to new images from different medical facilities and to new tasks in different settings. This rapid and generalisable learning ability is largely due to the compositional structure of image patterns in the human brain, which are not well represented in current medical models. In this paper, we study the utilisation of compositionality in learning more interpretable and generalisable representations for medical image segmentation. Overall, we propose that the underlying generative factors that are used to generate the medical images satisfy compositional equivariance property, where each factor is compositional (e.g. corresponds to human anatomy) and also equivariant to the task. Hence, a good representation that approximates well the ground truth factor has to be compositionally equivariant. By modelling the compositional representations with learnable von-Mises-Fisher (vMF) kernels, we explore how different design and learning biases can be used to enforce the representations to be more compositionally equivariant under un-, weakly-, and semi-supervised settings. Extensive results show that our methods achieve the best performance over several strong baselines on the task of semi-supervised domain-generalised medical image segmentation. Code will be made publicly available upon acceptance at https://github.com/vios-s.

PMID:38277249 | DOI:10.1109/TMI.2024.3358955

Rev Sci Instrum. 2024 Jan 1;95(1):015115. doi: 10.1063/5.0172368.

ABSTRACT

Machine learning provides increasingly reliable assistance for medical experts in diagnosing coronary heart disease. This study proposes a deep learning hybrid model based coronary heart disease (CAD) prediction method, which can significantly improve the prediction accuracy compared to traditional solutions. This research scheme is based on the data of 7291 patients and proposes a hybrid model, which uses two different deep neural network models and a recurrent neural network model as the main model for training. The prediction results based on the main model training use a k-nearest neighbor model for secondary training so as to improve the accuracy of coronary heart disease prediction. The comparison between the model prediction results and the clinical diagnostic results shows that the prediction model has a prediction accuracy rate of 82.8%, a prediction precision rate of 87.08%, a prediction recall rate of 88.57%, a prediction F1-score of 87.82%, and an area under the curve value of 0.8 in the test set. Compared to single model machine learning predictions, the hybrid model has a significantly improved accuracy and has effectively solved the problem of overfitting. A deep learning based CAD prediction hybrid model that combines multiple weak models into a strong model can fully explore the complex inter-relationships between various features under limited feature values and sample size, improve the evaluation indicators of the prediction model, and provide effective auxiliary support for CAD diagnosis.

PMID:38276898 | DOI:10.1063/5.0172368

Rev Sci Instrum. 2024 Jan 1;95(1):015113. doi: 10.1063/5.0172297.

ABSTRACT

Coordinated Universal Time (UTC), produced by the Bureau International des Poids et Mesures (BIPM), is the official worldwide time reference. Given that there is no physical signal associated with UTC, physical realizations of the UTC, called UTC(k), are very important for demanding applications such as global navigation satellite systems, communication networks, and national defense and security, among others. Therefore, the prediction of the time differences UTC-UTC(k) is important to maintain the accuracy and stability of the UTC(k) timescales. In this paper, we report for the first time the use of a deep learning (DL) technique called Gated Recurrent Unit (GRU) to predict a sequence of H futures values of the time differences UTC-UTC(k) for ten different UTC(k) timescales. UTC-UTC(k) time differences published on the monthly Circular T document of the BIPM are used as training samples. We utilize a multiple-input, multiple-output prediction strategy. After a training process where about 300 past values of the difference UTC-UTC(k) are used, H (H = 6) values of the Circular T can be predicted using p (typically p = 6) past values. The model has been tested with data from ten different UTC(k) timescales. When comparing GRU results with other standard DL algorithms, we found that the GRU approximation has a good performance in predicting UTC(k) timescales. According to our results, the GRU error in predicting UTC-UTC(k) values is typically 1 ns. The frequency instability of the UTC(k) timescale is the main limitation in reducing the GRU error in the time difference prediction.

PMID:38276897 | DOI:10.1063/5.0172297

Molecules. 2024 Jan 19;29(2):495. doi: 10.3390/molecules29020495.

ABSTRACT

In recent years, the application of deep learning in molecular de novo design has gained significant attention. One successful approach involves using SMILES representations of molecules and treating the generation task as a text generation problem, yielding promising results. However, the generation of more effective and novel molecules remains a key research area. Due to the fact that a molecule can have multiple SMILES representations, it is not sufficient to consider only one of them for molecular generation. To make up for this deficiency, and also motivated by the advancements in contrastive learning in natural language processing, we propose a contrastive learning framework called CONSMI to learn more comprehensive SMILES representations. This framework leverages different SMILES representations of the same molecule as positive examples and other SMILES representations as negative examples for contrastive learning. The experimental results of generation tasks demonstrate that CONSMI significantly enhances the novelty of generated molecules while maintaining a high validity. Moreover, the generated molecules have similar chemical properties compared to the original dataset. Additionally, we find that CONSMI can achieve favorable results in classifier tasks, such as the compound-protein interaction task.

PMID:38276573 | DOI:10.3390/molecules29020495

Pharmaceutics. 2023 Dec 19;16(1):3. doi: 10.3390/pharmaceutics16010003.

ABSTRACT

Patients with multiple sclerosis (MS) often take multiple drugs at the same time to modify the course of disease, alleviate neurological symptoms and manage co-existing conditions. A major consequence for a patient taking different medications is a higher risk of treatment failure and side effects. This is because a drug may alter the pharmacokinetic and/or pharmacodynamic properties of another drug, which is referred to as drug-drug interaction (DDI). We aimed to predict interactions of drugs that are used by patients with MS based on a deep neural network (DNN) using structural information as input. We further aimed to identify potential drug-food interactions (DFIs), which can affect drug efficacy and patient safety as well. We used DeepDDI, a multi-label classification model of specific DDI types, to predict changes in pharmacological effects and/or the risk of adverse drug events when two or more drugs are taken together. The original model with ~34 million trainable parameters was updated using >1 million DDIs recorded in the DrugBank database. Structure data of food components were obtained from the FooDB database. The medication plans of patients with MS (n = 627) were then searched for pairwise interactions between drug and food compounds. The updated DeepDDI model achieved accuracies of 92.2% and 92.1% on the validation and testing sets, respectively. The patients with MS used 312 different small molecule drugs as prescription or over-the-counter medications. In the medication plans, we identified 3748 DDIs in DrugBank and 13,365 DDIs using DeepDDI. At least one DDI was found for most patients (n = 509 or 81.2% based on the DNN model). The predictions revealed that many patients would be at increased risk of bleeding and bradycardic complications due to a potential DDI if they were to start a disease-modifying therapy with cladribine (n = 242 or 38.6%) and fingolimod (n = 279 or 44.5%), respectively. We also obtained numerous potential interactions for Bruton's tyrosine kinase inhibitors that are in clinical development for MS, such as evobrutinib (n = 434 DDIs). Food sources most often related to DFIs were corn (n = 5456 DFIs) and cow's milk (n = 4243 DFIs). We demonstrate that deep learning techniques can exploit chemical structure similarity to accurately predict DDIs and DFIs in patients with MS. Our study specifies drug pairs that potentially interact, suggests mechanisms causing adverse drug effects, informs about whether interacting drugs can be replaced with alternative drugs to avoid critical DDIs and provides dietary recommendations for MS patients who are taking certain drugs.

PMID:38276481 | DOI:10.3390/pharmaceutics16010003

Materials (Basel). 2024 Jan 21;17(2):510. doi: 10.3390/ma17020510.

ABSTRACT

In laser powder bed fusion processes, keyholes are the gaseous cavities formed where laser interacts with metal, and their morphologies play an important role in defect formation and the final product quality. The in-situ X-ray imaging technique can monitor the keyhole dynamics from the side and capture keyhole shapes in the X-ray image stream. Keyhole shapes in X-ray images are then often labeled by humans for analysis, which increasingly involves attempting to correlate keyhole shapes with defects using machine learning. However, such labeling is tedious, time-consuming, error-prone, and cannot be scaled to large data sets. To use keyhole shapes more readily as the input to machine learning methods, an automatic tool to identify keyhole regions is desirable. In this paper, a deep-learning-based computer vision tool that can automatically segment keyhole shapes out of X-ray images is presented. The pipeline contains a filtering method and an implementation of the BASNet deep learning model to semantically segment the keyhole morphologies out of X-ray images. The presented tool shows promising average accuracy of 91.24% for keyhole area, and 92.81% for boundary shape, for a range of test dataset conditions in Al6061 (and one AliSi10Mg) alloys, with 300 training images/labels and 100 testing images for each trial. Prospective users may apply the presently trained tool or a retrained version following the approach used here to automatically label keyhole shapes in large image sets.

PMID:38276449 | DOI:10.3390/ma17020510

Sensors (Basel). 2024 Jan 22;24(2):715. doi: 10.3390/s24020715.

ABSTRACT

The subtype diagnosis and severity classification of mood disorder have been made through the judgment of verified assistance tools and psychiatrists. Recently, however, many studies have been conducted using biomarker data collected from subjects to assist in diagnosis, and most studies use heart rate variability (HRV) data collected to understand the balance of the autonomic nervous system on statistical analysis methods to perform classification through statistical analysis. In this research, three mood disorder severity or subtype classification algorithms are presented through multimodal analysis of data on the collected heart-related data variables and hidden features from the variables of time and frequency domain of HRV. Comparing the classification performance of the statistical analysis widely used in existing major depressive disorder (MDD), anxiety disorder (AD), and bipolar disorder (BD) classification studies and the multimodality deep neural network analysis newly proposed in this study, it was confirmed that the severity or subtype classification accuracy performance of each disease improved by 0.118, 0.231, and 0.125 on average. Through the study, it was confirmed that deep learning analysis of biomarker data such as HRV can be applied as a primary identification and diagnosis aid for mental diseases, and that it can help to objectively diagnose psychiatrists in that it can confirm not only the diagnosed disease but also the current mood status.

PMID:38276406 | DOI:10.3390/s24020715

Sensors (Basel). 2024 Jan 22;24(2):701. doi: 10.3390/s24020701.

ABSTRACT

This paper introduces a new technique for automatic modulation classification (AMC) in Cognitive Radio (CR) networks. The method employs a straightforward classifier that utilizes high-order cumulant for training. It focuses on the statistical behavior of both analog modulation and digital schemes, which have received limited attention in previous works. The simulation results show that the proposed method performs well with different signal-to-noise ratios (SNRs) and channel conditions. The classifier's performance is superior to that of complex deep learning methods, making it suitable for deployment in CR networks' end units, especially in military and emergency service applications. The proposed method offers a cost-effective and high-quality solution for AMC that meets the strict demands of these critical applications.

PMID:38276393 | DOI:10.3390/s24020701

Sensors (Basel). 2024 Jan 22;24(2):702. doi: 10.3390/s24020702.

ABSTRACT

In recent years, optical camera communication (OCC) has garnered attention as a research focus. OCC uses optical light to transmit data by scattering the light in various directions. Although this can be advantageous with multiple transmitter scenarios, there are situations in which only a single transmitter is permitted to communicate. Therefore, this method is proposed to fulfill the latter requirement using 2D object size to calculate the proximity of the objects through an AI object detection model. This approach enables prioritization among transmitters based on the transmitter proximity to the receiver for communication, facilitating alternating communication with multiple transmitters. The image processing employed when receiving the signals from transmitters enables communication to be performed without the need to modify the camera parameters. During the implementation, the distance between the transmitter and receiver varied between 1.0 and 5.0 m, and the system demonstrated a maximum data rate of 3.945 kbps with a minimum BER of 4.2×10-3. Additionally, the system achieved high accuracy from the refined YOLOv8 detection algorithm, reaching 0.98 mAP at a 0.50 IoU.

PMID:38276392 | DOI:10.3390/s24020702

Sensors (Basel). 2024 Jan 22;24(2):693. doi: 10.3390/s24020693.

ABSTRACT

The term out-of-stock (OOS) describes a problem that occurs when shoppers come to a store and the product they are seeking is not present on its designated shelf. Missing products generate huge sales losses and may lead to a declining reputation or the loss of loyal customers. In this paper, we propose a novel deep-learning (DL)-based OOS-detection method that utilizes a two-stage training process and a post-processing technique designed for the removal of inaccurate detections. To develop the method, we utilized an OOS detection dataset that contains a commonly used fully empty OOS class and a novel class that represents the frontal OOS. We present a new image augmentation procedure in which some existing OOS instances are enlarged by duplicating and mirroring themselves over nearby products. An object-detection model is first pre-trained using only augmented shelf images and, then, fine-tuned on the original data. During the inference, the detected OOS instances are post-processed based on their aspect ratio. In particular, the detected instances are discarded if their aspect ratio is higher than the maximum or lower than the minimum instance aspect ratio found in the dataset. The experimental results showed that the proposed method outperforms the existing DL-based OOS-detection methods and detects fully empty and frontal OOS instances with 86.3% and 83.7% of the average precision, respectively.

PMID:38276384 | DOI:10.3390/s24020693

Sensors (Basel). 2024 Jan 21;24(2):682. doi: 10.3390/s24020682.

ABSTRACT

Early identification of respiratory irregularities is critical for improving lung health and reducing global mortality rates. The analysis of respiratory sounds plays a significant role in characterizing the respiratory system's condition and identifying abnormalities. The main contribution of this study is to investigate the performance when the input data, represented by cochleogram, is used to feed the Vision Transformer (ViT) architecture, since this input-classifier combination is the first time it has been applied to adventitious sound classification to our knowledge. Although ViT has shown promising results in audio classification tasks by applying self-attention to spectrogram patches, we extend this approach by applying the cochleogram, which captures specific spectro-temporal features of adventitious sounds. The proposed methodology is evaluated on the ICBHI dataset. We compare the classification performance of ViT with other state-of-the-art CNN approaches using spectrogram, Mel frequency cepstral coefficients, constant-Q transform, and cochleogram as input data. Our results confirm the superior classification performance combining cochleogram and ViT, highlighting the potential of ViT for reliable respiratory sound classification. This study contributes to the ongoing efforts in developing automatic intelligent techniques with the aim to significantly augment the speed and effectiveness of respiratory disease detection, thereby addressing a critical need in the medical field.

PMID:38276373 | DOI:10.3390/s24020682

Curr Med Chem. 2024 Jan 24. doi: 10.2174/0109298673264899231206093930. Online ahead of print.

ABSTRACT

The application of therapeutic peptides in clinical practice has significantly progressed in the past decades. However, immunogenicity remains an inevitable and crucial issue in the development of therapeutic peptides. The prediction of antigenic peptides presented by MHC class II is a critical approach to evaluating the immunogenicity of therapeutic peptides. With the continuous upgrade of algorithms and databases in recent years, the prediction accuracy has been significantly improved. This has made in silico evaluation an important component of immunogenicity assessment in therapeutic peptide development. In this review, we summarize the development of peptide-MHC-II binding prediction methods for antigenic peptides presented by MHC class II molecules and provide a systematic explanation of the most advanced ones, aiming to deepen our understanding of this field that requires particular attention.

PMID:38275064 | DOI:10.2174/0109298673264899231206093930

Comput Struct Biotechnol J. 2023 Dec 29;23:559-565. doi: 10.1016/j.csbj.2023.12.041. eCollection 2024 Dec.

ABSTRACT

Escherichia coli (E. coli) has become a particular concern due to the increasing incidence of antimicrobial resistance (AMR) observed worldwide. Using machine learning (ML) to predict E. coli AMR is a more efficient method than traditional laboratory testing. However, further improvement in the predictive performance of existing models remains challenging. In this study, we collected 1937 high-quality whole genome sequencing (WGS) data from public databases with an antimicrobial resistance phenotype and modified the existing workflow by adding an attention mechanism to enable the modified workflow to focus more on core single nucleotide polymorphisms (SNPs) that may significantly lead to the development of AMR in E. coli. While comparing the model performance before and after adding the attention mechanism, we also performed a cross-comparison among the published models using random forest (RF), support vector machine (SVM), logistic regression (LR), and convolutional neural network (CNN). Our study demonstrates that the discriminative positional colors of Chaos Game Representation (CGR) images can selectively influence and highlight genome regions without prior knowledge, enhancing prediction accuracy. Furthermore, we developed an online tool (https://github.com/tjiaa/E.coli-ML/tree/main) for assisting clinicians in the rapid prediction of the AMR phenotype of E. coli and accelerating clinical decision-making.

PMID:38274998 | PMC:PMC10809114 | DOI:10.1016/j.csbj.2023.12.041