A fluorescent sensor array based on antibiotic-stabilized metal nanoclusters for the multiplex detection of bacteria.

To address the need for facile, rapid detection of pathogens in water supplies, a fluorescent sensing array platform based on antibiotic-stabilized metal nanoclusters was developed for the multiplex detection of pathogens. Using five common antibiotics, eight different nanoclusters (NCs) were synthesized including ampicillin stabilized copper NCs, cefepime stabilized gold and copper NCs, kanamycin stabilized gold and copper NCs, lysozyme stabilized gold NCs, and vancomycin stabilized gold/silver and copper NCs. Based on the different interaction of each NC with the bacteria strains, unique patterns were generated. Various machine learning algorithms were employed for pattern discernment, among which the artificial neural networks proved to have the highest performance, with an accuracy of 100%. The developed prediction model performed well on an independent test dataset and on real samples gathered from drinking water, tap water and the Anzali Lagoon water, with prediction accuracy of 96.88% and 95.14%, respectively. This work demonstrates how generic antibiotics can be implemented for NC synthesis and used as recognition elements for pathogen detection. Furthermore, it displays how merging machine learning techniques can elevate sensitivity of analytical devices.

Forecasting vaping health risks through neural network model prediction of flavour pyrolysis reactions.

Vaping involves the heating of chemical solutions (e-liquids) to high temperatures prior to lung inhalation. A risk exists that these chemicals undergo thermal decomposition to new chemical entities, the composition and health implications of which are largely unknown. To address this concern, a graph-convolutional neural network (NN) model was used to predict pyrolysis reactivity of 180 e-liquid chemical flavours. The output of this supervised machine learning approach was a dataset of probability ranked pyrolysis transformations and their associated 7307 products. To refine this dataset, the molecular weight of each NN predicted product was automatically correlated with experimental mass spectrometry (MS) fragmentation data for each flavour chemical. This blending of deep learning methods with experimental MS data identified 1169 molecular weight matches that prioritized these compounds for further analysis. The average number of discrete matches per flavour between NN predictions and MS fragmentation was 6.4 with 92.8% of flavours having at least one match. Globally harmonized system classifications for NN/MS matches were extracted from PubChem, revealing that 127 acute toxic, 153 health hazard and 225 irritant classifications were predicted. This approach may reveal the longer-term health risks of vaping in advance of clinical diseases emerging in the general population.

ARGNet: using deep neural networks for robust identification and classification of antibiotic resistance genes from sequences.

BACKGROUND: Emergence of antibiotic resistance in bacteria is an important threat to global health. Antibiotic resistance genes (ARGs) are some of the key components to define bacterial resistance and their spread in different environments. Identification of ARGs, particularly from high-throughput sequencing data of the specimens, is the state-of-the-art method for comprehensively monitoring their spread and evolution. Current computational methods to identify ARGs mainly rely on alignment-based sequence similarities with known ARGs. Such approaches are limited by choice of reference databases and may potentially miss novel ARGs. The similarity thresholds are usually simple and could not accommodate variations across different gene families and regions. It is also difficult to scale up when sequence data are increasing. RESULTS: In this study, we developed ARGNet, a deep neural network that incorporates an unsupervised learning autoencoder model to identify ARGs and a multiclass classification convolutional neural network to classify ARGs that do not depend on sequence alignment. This approach enables a more efficient discovery of both known and novel ARGs. ARGNet accepts both amino acid and nucleotide sequences of variable lengths, from partial (30-50 aa; 100-150 nt) sequences to full-length protein or genes, allowing its application in both target sequencing and metagenomic sequencing. Our performance evaluation showed that ARGNet outperformed other deep learning models including DeepARG and HMD-ARG in most of the application scenarios especially quasi-negative test and the analysis of prediction consistency with phylogenetic tree. ARGNet has a reduced inference runtime by up to 57% relative to DeepARG. CONCLUSIONS: ARGNet is flexible, efficient, and accurate at predicting a broad range of ARGs from the sequencing data. ARGNet is freely available at https://github.com/id-bioinfo/ARGNet , with an online service provided at https://ARGNet.hku.hk . Video Abstract.

Interpretable (not just posthoc-explainable) medical claims modeling for discharge placement to reduce preventable all-cause readmissions or death.

We developed an inherently interpretable multilevel Bayesian framework for representing variation in regression coefficients that mimics the piecewise linearity of ReLU-activated deep neural networks. We used the framework to formulate a survival model for using medical claims to predict hospital readmission and death that focuses on discharge placement, adjusting for confounding in estimating causal local average treatment effects. We trained the model on a 5% sample of Medicare beneficiaries from 2008 and 2011, based on their 2009-2011 inpatient episodes (approximately 1.2 million), and then tested the model on 2012 episodes (approximately 400 thousand). The model scored an out-of-sample AUROC of approximately 0.75 on predicting all-cause readmissions-defined using official Centers for Medicare and Medicaid Services (CMS) methodology-or death within 30-days of discharge, being competitive against XGBoost and a Bayesian deep neural network, demonstrating that one need-not sacrifice interpretability for accuracy. Crucially, as a regression model, it provides what blackboxes cannot-its exact gold-standard global interpretation, explicitly defining how the model performs its internal "reasoning" for mapping the input data features to predictions. In doing so, we identify relative risk factors and quantify the effect of discharge placement. We also show that the posthoc explainer SHAP provides explanations that are inconsistent with the ground truth model reasoning that our model readily admits.

Dynamic multilayer growth: Parallel vs. sequential approaches.

The decision of when to add a new hidden unit or layer is a fundamental challenge for constructive algorithms. It becomes even more complex in the context of multiple hidden layers. Growing both network width and depth offers a robust framework for leveraging the ability to capture more information from the data and model more complex representations. In the context of multiple hidden layers, should growing units occur sequentially with hidden units only being grown in one layer at a time or in parallel with hidden units growing across multiple layers simultaneously? The effects of growing sequentially or in parallel are investigated using a population dynamics-inspired growing algorithm in a multilayer context. A modified version of the constructive growing algorithm capable of growing in parallel is presented. Sequential and parallel growth methodologies are compared in a three-hidden layer multilayer perceptron on several benchmark classification tasks. Several variants of these approaches are developed for a more in-depth comparison based on the type of hidden layer initialization and the weight update methods employed. Comparisons are then made to another sequential growing approach, Dynamic Node Creation. Growing hidden layers in parallel resulted in comparable or higher performances than sequential approaches. Growing hidden layers in parallel promotes growing narrower deep architectures tailored to the task. Dynamic growth inspired by population dynamics offers the potential to grow the width and depth of deeper neural networks in either a sequential or parallel fashion.

AMPActiPred: A three-stage framework for predicting antibacterial peptides and activity levels with deep forest.

The emergence and spread of antibiotic-resistant bacteria pose a significant public health threat, necessitating the exploration of alternative antibacterial strategies. Antibacterial peptide (ABP) is a kind of antimicrobial peptide (AMP) that has the potential ability to fight against bacteria infection, offering a promising avenue for developing novel therapeutic interventions. This study introduces AMPActiPred, a three-stage computational framework designed to identify ABPs, characterize their activity against diverse bacterial species, and predict their activity levels. AMPActiPred employed multiple effective peptide descriptors to effectively capture the compositional features and physicochemical properties of peptides. AMPActiPred utilized deep forest architecture, a cascading architecture similar to deep neural networks, capable of effectively processing and exploring original features to enhance predictive performance. In the first stage, AMPActiPred focuses on ABP identification, achieving an Accuracy of 87.6% and an MCC of 0.742 on an elaborate dataset, demonstrating state-of-the-art performance. In the second stage, AMPActiPred achieved an average GMean at 82.8% in identifying ABPs targeting 10 bacterial species, indicating AMPActiPred can achieve balanced predictions regarding the functional activity of ABP across this set of species. In the third stage, AMPActiPred demonstrates robust predictive capabilities for ABP activity levels with an average PCC of 0.722. Furthermore, AMPActiPred exhibits excellent interpretability, elucidating crucial features associated with antibacterial activity. AMPActiPred is the first computational framework capable of predicting targets and activity levels of ABPs. Finally, to facilitate the utilization of AMPActiPred, we have established a user-friendly web interface deployed at https://awi.cuhk.edu.cn/∼AMPActiPred/.

Diagnostic biomarker discovery from brain EEG data using LSTM, reservoir-SNN, and NeuCube methods in a pilot study comparing epilepsy and migraine.

The study introduces a new online spike encoding algorithm for spiking neural networks (SNN) and suggests new methods for learning and identifying diagnostic biomarkers using three prominent deep learning neural network models: deep BiLSTM, reservoir SNN, and NeuCube. EEG data from datasets related to epilepsy, migraine, and healthy subjects are employed. Results reveal that BiLSTM hidden neurons capture biological significance, while reservoir SNN activities and NeuCube spiking dynamics identify EEG channels as diagnostic biomarkers. BiLSTM and reservoir SNN achieve 90 and 85% classification accuracy, while NeuCube achieves 97%, all methods pinpointing potential biomarkers like T6, F7, C4, and F8. The research bears implications for refining online EEG classification, analysis, and early brain state diagnosis, enhancing AI models with interpretability and discovery. The proposed techniques hold promise for streamlined brain-computer interfaces and clinical applications, representing a significant advancement in pattern discovery across the three most popular neural network methods for addressing a crucial problem. Further research is planned to study how early can these diagnostic biomarkers predict an onset of brain states.

GNNGL-PPI: multi-category prediction of protein-protein interactions using graph neural networks based on global graphs and local subgraphs.

Most proteins exert their functions by interacting with other proteins, making the identification of protein-protein interactions (PPI) crucial for understanding biological activities, pathological mechanisms, and clinical therapies. Developing effective and reliable computational methods for predicting PPI can significantly reduce the time-consuming and labor-intensive associated traditional biological experiments. However, accurately identifying the specific categories of protein-protein interactions and improving the prediction accuracy of the computational methods remain dual challenges. To tackle these challenges, we proposed a novel graph neural network method called GNNGL-PPI for multi-category prediction of PPI based on global graphs and local subgraphs. GNNGL-PPI consisted of two main components: using Graph Isomorphism Network (GIN) to extract global graph features from PPI network graph, and employing GIN As Kernel (GIN-AK) to extract local subgraph features from the subgraphs of protein vertices. Additionally, considering the imbalanced distribution of samples in each category within the benchmark datasets, we introduced an Asymmetric Loss (ASL) function to further enhance the predictive performance of the method. Through evaluations on six benchmark test sets formed by three different dataset partitioning algorithms (Random, BFS, DFS), GNNGL-PPI outperformed the state-of-the-art multi-category prediction methods of PPI, as measured by the comprehensive performance evaluation metric F1-measure. Furthermore, interpretability analysis confirmed the effectiveness of GNNGL-PPI as a reliable multi-category prediction method for predicting protein-protein interactions.

Clinical application of high-resolution spiral CT scanning in the diagnosis of auriculotemporal and ossicle.

Precision and intelligence in evaluating the complexities of middle ear structures are required to diagnose auriculotemporal and ossicle-related diseases within otolaryngology. Due to the complexity of the anatomical details and the varied etiologies of illnesses such as trauma, chronic otitis media, and congenital anomalies, traditional diagnostic procedures may not yield accurate diagnoses. This research intends to enhance the diagnosis of diseases of the auriculotemporal region and ossicles by combining High-Resolution Spiral Computed Tomography (HRSCT) scanning with Deep Learning Techniques (DLT). This study employs a deep learning method, Convolutional Neural Network-UNet (CNN-UNet), to extract sub-pixel information from medical photos. This method equips doctors and researchers with cutting-edge resources, leading to groundbreaking discoveries and better patient healthcare. The research effort is the interaction between the CNN-UNet model and high-resolution Computed Tomography (CT) scans, automating activities including ossicle segmentation, fracture detection, and disruption cause classification, accelerating the diagnostic process and increasing clinical decision-making. The suggested HRSCT-DLT model represents the integration of high-resolution spiral CT scans with the CNN-UNet model, which has been fine-tuned to address the nuances of auriculotemporal and ossicular diseases. This novel combination improves diagnostic efficiency and our overall understanding of these intricate diseases. The results of this study highlight the promise of combining high-resolution CT scanning with the CNN-UNet model in otolaryngology, paving the way for more accurate diagnosis and more individualized treatment plans for patients experiencing auriculotemporal and ossicle-related disruptions.

DCGAN-DTA: Predicting drug-target binding affinity with deep convolutional generative adversarial networks.

BACKGROUND: In recent years, there has been a growing interest in utilizing computational approaches to predict drug-target binding affinity, aiming to expedite the early drug discovery process. To address the limitations of experimental methods, such as cost and time, several machine learning-based techniques have been developed. However, these methods encounter certain challenges, including the limited availability of training data, reliance on human intervention for feature selection and engineering, and a lack of validation approaches for robust evaluation in real-life applications. RESULTS: To mitigate these limitations, in this study, we propose a method for drug-target binding affinity prediction based on deep convolutional generative adversarial networks. Additionally, we conducted a series of validation experiments and implemented adversarial control experiments using straw models. These experiments serve to demonstrate the robustness and efficacy of our predictive models. We conducted a comprehensive evaluation of our method by comparing it to baselines and state-of-the-art methods. Two recently updated datasets, namely the BindingDB and PDBBind, were used for this purpose. Our findings indicate that our method outperforms the alternative methods in terms of three performance measures when using warm-start data splitting settings. Moreover, when considering physiochemical-based cold-start data splitting settings, our method demonstrates superior predictive performance, particularly in terms of the concordance index. CONCLUSION: The results of our study affirm the practical value of our method and its superiority over alternative approaches in predicting drug-target binding affinity across multiple validation sets. This highlights the potential of our approach in accelerating drug repurposing efforts, facilitating novel drug discovery, and ultimately enhancing disease treatment. The data and source code for this study were deposited in the GitHub repository, https://github.com/mojtabaze7/DCGAN-DTA . Furthermore, the web server for our method is accessible at https://dcgan.shinyapps.io/bindingaffinity/ .

Prediction of anticancer drug sensitivity using an interpretable model guided by deep learning.

BACKGROUND: The prediction of drug sensitivity plays a crucial role in improving the therapeutic effect of drugs. However, testing the effectiveness of drugs is challenging due to the complex mechanism of drug reactions and the lack of interpretability in most machine learning and deep learning methods. Therefore, it is imperative to establish an interpretable model that receives various cell line and drug feature data to learn drug response mechanisms and achieve stable predictions between available datasets. RESULTS: This study proposes a new and interpretable deep learning model, DrugGene, which integrates gene expression, gene mutation, gene copy number variation of cancer cells, and chemical characteristics of anticancer drugs to predict their sensitivity. This model comprises two different branches of neural networks, where the first involves a hierarchical structure of biological subsystems that uses the biological processes of human cells to form a visual neural network (VNN) and an interpretable deep neural network for human cancer cells. DrugGene receives genotype input from the cell line and detects changes in the subsystem states. We also employ a traditional artificial neural network (ANN) to capture the chemical structural features of drugs. DrugGene generates final drug response predictions by combining VNN and ANN and integrating their outputs into a fully connected layer. The experimental results using drug sensitivity data extracted from the Cancer Drug Sensitivity Genome Database and the Cancer Treatment Response Portal v2 reveal that the proposed model is better than existing prediction methods. Therefore, our model achieves higher accuracy, learns the reaction mechanisms between anticancer drugs and cell lines from various features, and interprets the model's predicted results. CONCLUSIONS: Our method utilizes biological pathways to construct neural networks, which can use genotypes to monitor changes in the state of network subsystems, thereby interpreting the prediction results in the model and achieving satisfactory prediction accuracy. This will help explore new directions in cancer treatment. More available code resources can be downloaded for free from GitHub ( https://github.com/pangweixiong/DrugGene ).

TransPTM: a transformer-based model for non-histone acetylation site prediction.

Protein acetylation is one of the extensively studied post-translational modifications (PTMs) due to its significant roles across a myriad of biological processes. Although many computational tools for acetylation site identification have been developed, there is a lack of benchmark dataset and bespoke predictors for non-histone acetylation site prediction. To address these problems, we have contributed to both dataset creation and predictor benchmark in this study. First, we construct a non-histone acetylation site benchmark dataset, namely NHAC, which includes 11 subsets according to the sequence length ranging from 11 to 61 amino acids. There are totally 886 positive samples and 4707 negative samples for each sequence length. Secondly, we propose TransPTM, a transformer-based neural network model for non-histone acetylation site predication. During the data representation phase, per-residue contextualized embeddings are extracted using ProtT5 (an existing pre-trained protein language model). This is followed by the implementation of a graph neural network framework, which consists of three TransformerConv layers for feature extraction and a multilayer perceptron module for classification. The benchmark results reflect that TransPTM has the competitive performance for non-histone acetylation site prediction over three state-of-the-art tools. It improves our comprehension on the PTM mechanism and provides a theoretical basis for developing drug targets for diseases. Moreover, the created PTM datasets fills the gap in non-histone acetylation site datasets and is beneficial to the related communities. The related source code and data utilized by TransPTM are accessible at https://www.github.com/TransPTM/TransPTM.

BACK-to-MOVE: Machine learning and computer vision model automating clinical classification of non-specific low back pain for personalised management.

BACKGROUND: Low back pain (LBP) is a major global disability contributor with profound health and socio-economic implications. The predominant form is non-specific LBP (NSLBP), lacking treatable pathology. Active physical interventions tailored to individual needs and capabilities are crucial for its management. However, the intricate nature of NSLBP and complexity of clinical classification systems necessitating extensive clinical training, hinder customised treatment access. Recent advancements in machine learning and computer vision demonstrate promise in characterising NSLBP altered movement patters through wearable sensors and optical motion capture. This study aimed to develop and evaluate a machine learning model (i.e., 'BACK-to-MOVE') for NSLBP classification trained with expert clinical classification, spinal motion data from a standard video alongside patient-reported outcome measures (PROMs). METHODS: Synchronised video and three-dimensional (3D) motion data was collected during forward spinal flexion from 83 NSLBP patients. Two physiotherapists independently classified them as motor control impairment (MCI) or movement impairment (MI), with conflicts resolved by a third expert. The Convolutional Neural Networks (CNNs) architecture, HigherHRNet, was chosen for effective pose estimation from video data. The model was validated against 3D motion data (subset of 62) and trained on the freely available MS-COCO dataset for feature extraction. The Back-to-Move classifier underwent fine-tuning through feed-forward neural networks using labelled examples from the training dataset. Evaluation utilised 5-fold cross-validation to assess accuracy, specificity, sensitivity, and F1 measure. RESULTS: Pose estimation's Mean Square Error of 0.35 degrees against 3D motion data demonstrated strong criterion validity. Back-to-Move proficiently differentiated MI and MCI classes, yielding 93.98% accuracy, 96.49% sensitivity (MI detection), 88.46% specificity (MCI detection), and an F1 measure of .957. Incorporating PROMs curtailed classifier performance (accuracy: 68.67%, sensitivity: 91.23%, specificity: 18.52%, F1: .800). CONCLUSION: This study is the first to demonstrate automated clinical classification of NSLBP using computer vision and machine learning with standard video data, achieving accuracy comparable to expert consensus. Automated classification of NSLBP based on altered movement patters video-recorded during routine clinical examination could expedite personalised NSLBP rehabilitation management, circumventing existing healthcare constraints. This advancement holds significant promise for patients and healthcare services alike.

Explainable machine learning approach for cancer prediction through binarilization of RNA sequencing data.

In recent years, researchers have proven the effectiveness and speediness of machine learning-based cancer diagnosis models. However, it is difficult to explain the results generated by machine learning models, especially ones that utilized complex high-dimensional data like RNA sequencing data. In this study, we propose the binarilization technique as a novel way to treat RNA sequencing data and used it to construct explainable cancer prediction models. We tested our proposed data processing technique on five different models, namely neural network, random forest, xgboost, support vector machine, and decision tree, using four cancer datasets collected from the National Cancer Institute Genomic Data Commons. Since our datasets are imbalanced, we evaluated the performance of all models using metrics designed for imbalance performance like geometric mean, Matthews correlation coefficient, F-Measure, and area under the receiver operating characteristic curve. Our approach showed comparative performance while relying on less features. Additionally, we demonstrated that data binarilization offers higher explainability by revealing how each feature affects the prediction. These results demonstrate the potential of data binarilization technique in improving the performance and explainability of RNA sequencing based cancer prediction models.

Short-term prediction of PM2.5 concentration by hybrid neural network based on sequence decomposition.

Accurate forecasting of PM2.5 concentrations serves as a critical tool for mitigating air pollution. This study introduces a novel hybrid prediction model, termed MIC-CEEMDAN-CNN-BiGRU, for short-term forecasting of PM2.5 concentrations using a 24-hour historical data window. Utilizing the Maximal Information Coefficient (MIC) for feature selection, the model integrates Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Convolutional Neural Network (CNN), and Bidirectional Recurrent Gated Neural Network (BiGRU) to optimize predictive accuracy. We used 2016 PM2.5 monitoring data from Beijing, China as the empirical basis of this study and compared the model with several deep learning frameworks. RNN, LSTM, GRU, and other hybrid models based on GRU, respectively. The experimental results show that the prediction results of the hybrid model proposed in this question are more accurate than those of other models, and the R2 of the hybrid model proposed in this paper improves the R2 by nearly 5 percentage points compared with that of the single model; reduces the MAE by nearly 5 percentage points; and reduces the RMSE by nearly 11 percentage points. The results show that the hybrid prediction model proposed in this study is more accurate than other models in predicting PM2.5.

A dual-branch selective attention capsule network for classifying kiwifruit soft rot with hyperspectral images.

Kiwifruit soft rot is highly contagious and causes serious economic loss. Therefore, early detection and elimination of soft rot are important for postharvest treatment and storage of kiwifruit. This study aims to accurately detect kiwifruit soft rot based on hyperspectral images by using a deep learning approach for image classification. A dual-branch selective attention capsule network (DBSACaps) was proposed to improve the classification accuracy. The network uses two branches to separately extract the spectral and spatial features so as to reduce their mutual interference, followed by fusion of the two features through the attention mechanism. Capsule network was used instead of convolutional neural networks to extract the features and complete the classification. Compared with existing methods, the proposed method exhibited the best classification performance on the kiwifruit soft rot dataset, with an overall accuracy of 97.08% and a 97.83% accuracy for soft rot. Our results confirm that potential soft rot of kiwifruit can be detected using hyperspectral images, which may contribute to the construction of smart agriculture.

Fine tuning deep learning models for breast tumor classification.

This paper proposes an approach to enhance the differentiation task between benign and malignant Breast Tumors (BT) using histopathology images from the BreakHis dataset. The main stages involve preprocessing, which encompasses image resizing, data partitioning (training and testing sets), followed by data augmentation techniques. Both feature extraction and classification tasks are employed by a Custom CNN. The experimental results show that the proposed approach using the Custom CNN model exhibits better performance with an accuracy of 84% than applying the same approach using other pretrained models, including MobileNetV3, EfficientNetB0, Vgg16, and ResNet50V2, that present relatively lower accuracies, ranging from 74 to 82%; these four models are used as both feature extractors and classifiers. To increase the accuracy and other performance metrics, Grey Wolf Optimization (GWO), and Modified Gorilla Troops Optimization (MGTO) metaheuristic optimizers are applied to each model separately for hyperparameter tuning. In this case, the experimental results show that the Custom CNN model, refined with MGTO optimization, reaches an exceptional accuracy of 93.13% in just 10 iterations, outperforming the other state-of-the-art methods, and the other four used pretrained models based on the BreakHis dataset.

Distracted on duty: A theory-based exploration of influences leading to mobile phone distracted riding among food delivery workers.

Using mobile phones while riding is a form of distracted riding that significantly elevates crash risk. Regrettably, the factors contributing to mobile phone use while riding (MPUWR) among food delivery riders remain under-researched. Addressing this literature gap, the current study employs the Job Demands-Resources (JD-R) model and various socio-economic factors to examine the determinants of MPUWR. The research incorporates data from 558 delivery workers in Hanoi and Ho Chi Minh City, Vietnam. The study utilizes two analytical methods to empirically test the hypotheses, considering non-linear relationships between variables: Partial Least Square Structural Equation Modelling (PLS-SEM) and Artificial Neural Network (ANN). The results reveal mixed impacts of factors connected to job resources. Although social support appears to deter MPUWR, work autonomy and rewards seemingly encourage it. Furthermore, a predisposition towards risk-taking behaviour significantly impacts the frequency of mobile phone usage among delivery riders. Interestingly, riders with higher incomes and those who have previously been fined by the police exhibit more frequent mobile phone use. The findings of this study present valuable insights into the crucial factors to be addressed when designing interventions aimed at reducing phone use among food delivery riders.

Comprehensive analysis of clinical images contributions for melanoma classification using convolutional neural networks.

BACKGROUND: Timely diagnosis plays a critical role in determining melanoma prognosis, prompting the development of deep learning models to aid clinicians. Questions persist regarding the efficacy of clinical images alone or in conjunction with dermoscopy images for model training. This study aims to compare the classification performance for melanoma of three types of CNN models: those trained on clinical images, dermoscopy images, and a combination of paired clinical and dermoscopy images from the same lesion. MATERIALS AND METHODS: We divided 914 image pairs into training, validation, and test sets. Models were built using pre-trained Inception-ResNetV2 convolutional layers for feature extraction, followed by binary classification. Training comprised 20 models per CNN type using sets of random hyperparameters. Best models were chosen based on validation AUC-ROC. RESULTS: Significant AUC-ROC differences were found between clinical versus dermoscopy models (0.661 vs. 0.869, p < 0.001) and clinical versus clinical + dermoscopy models (0.661 vs. 0.822, p = 0.001). Significant sensitivity differences were found between clinical and dermoscopy models (0.513 vs. 0.799, p = 0.01), dermoscopy versus clinical + dermoscopy models (0.799 vs. 1.000, p = 0.02), and clinical versus clinical + dermoscopy models (0.513 vs. 1.000, p < 0.001). Significant specificity differences were found between dermoscopy versus clinical + dermoscopy models (0.800 vs. 0.288, p < 0.001) and clinical versus clinical + dermoscopy models (0.650 vs. 0.288, p < 0.001). CONCLUSION: CNN models trained on dermoscopy images outperformed those relying solely on clinical images under our study conditions. The potential advantages of incorporating paired clinical and dermoscopy images for CNN-based melanoma classification appear less clear based on our findings.

Neurotransmitter classification from electron microscopy images at synaptic sites in Drosophila melanogaster.

High-resolution electron microscopy of nervous systems has enabled the reconstruction of synaptic connectomes. However, we do not know the synaptic sign for each connection (i.e., whether a connection is excitatory or inhibitory), which is implied by the released transmitter. We demonstrate that artificial neural networks can predict transmitter types for presynapses from electron micrographs: a network trained to predict six transmitters (acetylcholine, glutamate, GABA, serotonin, dopamine, octopamine) achieves an accuracy of 87% for individual synapses, 94% for neurons, and 91% for known cell types across a D. melanogaster whole brain. We visualize the ultrastructural features used for prediction, discovering subtle but significant differences between transmitter phenotypes. We also analyze transmitter distributions across the brain and find that neurons that develop together largely express only one fast-acting transmitter (acetylcholine, glutamate, or GABA). We hope that our publicly available predictions act as an accelerant for neuroscientific hypothesis generation for the fly.