Computer methods and programs in biomedicine update最新文献

{"title":"Integrative in Silico modeling for mTOR inhibition: From ridge classifiers to descriptor-free deep neural networks","authors":"Seyed Alireza Khanghahi ,&nbsp;Hadi Kamkar ,&nbsp;Seyedehsamaneh Shojaeilangari ,&nbsp;Abdollah Allahverdi ,&nbsp;Parviz Abdolmaleki","doi":"10.1016/j.cmpbup.2025.100208","DOIUrl":"10.1016/j.cmpbup.2025.100208","url":null,"abstract":"<div><div>Inhibiting the mammalian Target of Rapamycin (mTOR) represents a promising strategy in cancer therapy due to its crucial role in cell growth, survival, and metabolism. Using a variety of quantitative structure-activity relationship (QSAR) models, we present a comprehensive comparison of deep learning (DL) and classical machine learning (ML) techniques for modeling mTOR inhibitor activity. Unlike prior studies that focused on specific algorithms or limited descriptors, we benchmark a wide range of models, from traditional models like Random Forest, logistic regression, and SVM to modern algorithms like CNNs, GRUs, and LSTMs, on both descriptor-based features obtained by Dragon and descriptor-free inputs, including raw SMILES string, and Morgan fingerprints. This comprehensive analysis provides a robust foundation for using the best QSAR models specific to mTOR inhibition. Our findings revealed that while the Random Forest classifier achieved the highest accuracy among all models (0.9290 accuracy, 0.8940 F1-score, 0.9737 AUC), DL methods also demonstrated strong predictive capabilities, with nearly all models attaining an accuracy above 0.90. Among the DL models, CNN-QSAR using Morgan fingerprints achieved the highest accuracy (0.9271), F1-score (0.8950), and AUC (0.9696), demonstrating its effectiveness in capturing structural characteristics. The GRU-QSAR and LSTM-QSAR models, which utilized tokenized SMILES, achieved accuracies of 0.9002 and 0.9021, F1-scores of 0.8595 and 0.8603, and AUCs of 0.9270 and 0.9529, respectively, leveraging their ability to process sequential data.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"8 ","pages":"Article 100208"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144653529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing diabetes prediction performance using feature selection based on grey wolf optimizer with autophagy mechanism","authors":"Sirmayanti ,&nbsp;Pulung Hendro Prastyo ,&nbsp;Mahyati","doi":"10.1016/j.cmpbup.2025.100207","DOIUrl":"10.1016/j.cmpbup.2025.100207","url":null,"abstract":"<div><div>Diabetes mellitus, often called a silent killer, is a chronic condition characterized by insufficient insulin production and elevated blood sugar levels, leading to complications in vital organs such as the nerves, eyes, and kidneys. Machine learning is a powerful tool for predicting diabetes; however, noisy features can negatively impact its accuracy, making an effective feature selection essential. This study proposes an improved feature selection approach for diabetes prediction, leveraging the Grey Wolf Optimizer with an integrated Autophagy Mechanism (GWO-AM) on the Pima Indian Diabetes Dataset. The autophagy mechanism, inspired by cellular self-degradation and recycling, is incorporated into GWO to enhance exploration and exploitation. The method was also tested on glioma and lung cancer datasets to assess scalability. Comprehensive experiments demonstrate that GWO-AM significantly improves prediction accuracy while reducing the number of selected features. For the diabetes dataset, GWO-AM achieved an accuracy of 90.91 %, outperforming existing methods. It also excelled in the glioma and lung cancer datasets, highlighting its potential for application to other medical datasets.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"8 ","pages":"Article 100207"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144680167","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimized soft-voting CNN ensemble using particle swarm optimization for endometrial cancer histopathology classification","authors":"Firas Ibrahim AlZobi ,&nbsp;Khalid Mansour ,&nbsp;Ahmad Nasayreh ,&nbsp;Ghassan Samara ,&nbsp;Neda’a Alsalman ,&nbsp;Ayah Bashkami ,&nbsp;Aseel Smerat ,&nbsp;Khalid M.O. Nahar","doi":"10.1016/j.cmpbup.2025.100217","DOIUrl":"10.1016/j.cmpbup.2025.100217","url":null,"abstract":"<div><div>The heterogeneity of endometrial cancer tissue presents a significant obstacle to accurate automated classification using histopathological images. While ensemble methods are a promising alternative to single Convolutional Neural Networks (CNNs), we introduce PSO-SV (Particle Swarm Optimization–Soft Voting), a novel framework that adaptively fuses the outputs of MobileNetV2, VGG19, DenseNet121, Swin Transformer, and Vision Transformer (ViT). Our key innovation is the use of Particle Swarm Optimization to dynamically determine the optimal contribution of each model in a soft-voting ensemble. We validated PSO-SV on two datasets, the first one consists from 11,977 tiles from 95 whole-slide images (WSIs) obtained from The Cancer Genome Atlas Uterine Corpus Endometrial Carcinoma (TCGA-UCEC) project, the other dataset consists of 3,302 images from 498 patients, which are categorized into four classes. The proposed framework achieved outstanding results, including 99.67% accuracy, a 99.67% F1-score, and an Area Under the Curve (AUC) of 99.9% on the first dataset and 99% for all metrics for the second dataset. It consistently outperformed all three individual CNNs and a traditional hard-voting ensemble, highlighting its ability to synergistically combine complementary model strengths. The PSO-SV framework offers a powerful and clinically promising approach for robust endometrial cancer classification.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"8 ","pages":"Article 100217"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144912933","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Time series analysis and prediction of the trends of COVID-19 epidemic in Singapore based on machine learning","authors":"Wenbin Yang ,&nbsp;Xin Chang","doi":"10.1016/j.cmpbup.2025.100190","DOIUrl":"10.1016/j.cmpbup.2025.100190","url":null,"abstract":"<div><div>The COVID-19 pandemic has posed a significant threat to global health, with ongoing rises in new cases and deaths in Singapore, profoundly affecting public health, social activities, and the economy. This study compares the performance of LSTM, GRU, and a composite prediction model (LSTM-GRU) using a dataset of new and cumulative COVID-19 cases in Singapore, provided by the World Health Organization. The analysis uses weekly cumulative data from 2020 to January 21, 2024, to forecast new cases for the upcoming weeks. Model performance is evaluated using RMSE, MAE, MAPE, and R2. The results show that the LSTM model outperforms others, particularly in capturing significant data fluctuations. This research provides insights into the trends of the pandemic in Singapore and offers a basis for further epidemiological control efforts in the region.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"7 ","pages":"Article 100190"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143783148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multipath2.0: Extending Multilayer Reproducible Pathway Models with Omics Data","authors":"Zaynab Hammoud ,&nbsp;Mohammad Al Maaz ,&nbsp;Alicia D'Angelo ,&nbsp;Frank Kramer","doi":"10.1016/j.cmpbup.2025.100189","DOIUrl":"10.1016/j.cmpbup.2025.100189","url":null,"abstract":"<div><h3>Background</h3><div>Biological systems are often perceived as independent and consequently analyzed individually. In the field of omics, multiple disciplines target the study of specific types of molecules, such as genomics. The support of more data sources in these analyses is becoming more crucial for understanding the interplay of biological systems. However, this requires integration of heterogeneous knowledge, which is considered highly challenging in bioinformatics and biomedicine. Therefore, the R package Multipath was developed to model biological pathways as multilayered graphs and integrate influencing knowledge including proteins and drugs. In its previous form, Multipath generated multilayer models of BioPAX-encoded pathways and included features to integrate drug and protein information from DrugBank and UniProtKB respectively. Although the model showed remarkable utility, including further data sources ensures enriching and expanding its capabilities.</div></div><div><h3>Results</h3><div>In this paper, a new version Multipath 2.0 is presented. The update additionally supports the two databases KEGG Genes and OMIM, which serve as the source for gene and disease entries and interactions. Information on the interactions between the previously and newly added nodes are extracted and integrated. The Multipath 2.0 offers features to update the original multilayer model and integrate the corresponding nodes and edges into two additional layers referring to KEGG Genes and OMIM. Furthermore, the embedded nodes are inter- and intra-connected using interactions from the original and newly supported data sources.</div></div><div><h3>Conclusion</h3><div>The R Package Multipath is presented with the main functions that are newly developed to support the integration of the databases KEGG Genes and OMIM. The model comprises multiple information relevant to the analysis of pathway data, and offers a reproducible and simplified view of complex, intertwined systems. Through the application of such highly integrated models the inference of new knowledge becomes easier and contributes to many fields such as drug repurposing and biomarker discovery.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"7 ","pages":"Article 100189"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143706495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A comprehensive review of the use of Shapley value to assess node importance in the analysis of biological networks","authors":"Giang Pham,&nbsp;Paolo Milazzo","doi":"10.1016/j.cmpbup.2025.100185","DOIUrl":"10.1016/j.cmpbup.2025.100185","url":null,"abstract":"<div><h3>Background:</h3><div>In 2017, Lundberg and Lee introduced SHAP, a breakthrough in Explainable AI, creatively applying the Shapley value to estimate the importance of input features in machine learning outputs. The Shapley value, from cooperative game theory, fairly distributes system gains among participants. Inspired by SHAP’s success, this survey explores the application of Shapley value-based methods in biological network analysis.</div></div><div><h3>Method:</h3><div>We conducted a comprehensive literature search on the application of the Shapley value in biological network analysis from 2004 to 2024. From this, we focused on studies that applied the Shapley value in innovative and non-trivial ways, distinct from its typical usage.</div></div><div><h3>Result:</h3><div>The review identified six original studies that provide novel applications of the Shapley value in analyzing biological networks. These methods have also inspired further development and applications. For each, we discuss the foundational contributions, subsequent advancements, and applications.</div></div><div><h3>Discussion:</h3><div>Although the reviewed methods share the common objective of using the Shapley value to estimate an element’s contribution within a system, each one takes a distinct approach to modeling the cooperative game. Some methods employ game settings that enable more efficient Shapley value calculations, albeit with a narrower scope, as they are tailored to specific problems. Other methods offer broader applicability but encounter the usual computational challenges associated with calculating the exact Shapley value due to its time complexity. Fortunately, these challenges can be mitigated through the use of approximation techniques. Despite the computational challenges, Shapley value-based methods demonstrate to be beneficial for the interpretation of biological networks.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"7 ","pages":"Article 100185"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143510601","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diagnosis of Alzheimer's disease using non-linear features of ERP signals through a hybrid attention-based CNN-LSTM model","authors":"Elias Mazrooei Rad ,&nbsp;Sayyed Majid Mazinani ,&nbsp;Seyyed Ali Zendehbad","doi":"10.1016/j.cmpbup.2025.100192","DOIUrl":"10.1016/j.cmpbup.2025.100192","url":null,"abstract":"<div><div>Biological signals have a dynamic and non-linear nature, and hence nonlinear analysis is important for understanding the signals. In this study, a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) model is proposed for the diagnosis of Alzheimer’s disease (AD) from the Event-Related Potential (ERP) signals obtained from the Electroencephalogram (EEG) data. The P300 component of the ERP signal, derived from acoustic stimulation, is a key indicator of AD, and its amplitude and latency are characterized. By using nonlinear features such as phase diagrams, correlation dimension, entropy, and Lyapunov exponents, the proposed model classifies AD stages. The hybrid CNN-LSTM architecture, enhanced by an attention mechanism, captures both spatial and temporal dependencies in the ERP signals, achieving high accuracy: For healthy people, 95 %, for mild AD patients, 92.5 %, and for severe AD patients, 97.5 %. The model achieves 75 % accuracy in recall mode for healthy individuals, 72.5 % for mild AD, and 87.5 % for severe AD. Results show that the proposed model outperforms traditional methods and provides a robust and accurate diagnostic framework for AD. The result of this approach is to show that the combination of non-linear EEG analysis with advanced deep learning methods could provide early and precise AD detection.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"7 ","pages":"Article 100192"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144089131","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Efficient synthesis of 3D MR images for schizophrenia diagnosis classification with generative adversarial networks","authors":"Sebastian King ,&nbsp;Yasmin Hollenbenders ,&nbsp;Alexandra Reichenbach","doi":"10.1016/j.cmpbup.2025.100197","DOIUrl":"10.1016/j.cmpbup.2025.100197","url":null,"abstract":"<div><div>Schizophrenia and other psychiatric disorders can greatly benefit from objective decision support in diagnosis and therapy. Machine learning approaches based on neuroimaging, e.g. magnetic resonance imaging (MRI), have the potential to serve this purpose. However, the medical data sets these algorithms can be trained on are often rather small, leading to overfit, and the resulting models can therewith not be transferred into a clinical setting. The generation of synthetic images from real data is a promising approach to overcome this shortcoming. Due to the small data set size and the size and complexity of medical images, i.e. their three-dimensional nature, those algorithms are challenged on several levels. We develop four generative adversarial network (GAN) architectures that tackle these challenges and evaluate them systematically with a data set of 193 MR images of schizophrenia patients and healthy controls. The best architecture, a GAN with spectral normalization regulation and an additional encoder (α-SN-GAN), is then extended with an auxiliary classifier into an ensemble of networks capable of generating distinct image sets for the two diagnostic categories. The synthetic images increase the accuracy of a diagnostic classifier from a baseline accuracy of around 61 % to 79 %. This novel end-to-end pipeline for schizophrenia diagnosis demonstrates a data and memory efficient approach to support clinical decision-making that can also be transferred to support other psychiatric disorders.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"8 ","pages":"Article 100197"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144322680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fostering digital health literacy to enhance trust and improve health outcomes","authors":"Kristine Sørensen","doi":"10.1016/j.cmpbup.2024.100140","DOIUrl":"https://doi.org/10.1016/j.cmpbup.2024.100140","url":null,"abstract":"","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"34 14","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139814804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning based detection of silicosis from computed tomography images","authors":"Hamit Aksoy ,&nbsp;Ümit Atila ,&nbsp;Sertaç Arslan","doi":"10.1016/j.cmpbup.2024.100166","DOIUrl":"10.1016/j.cmpbup.2024.100166","url":null,"abstract":"<div><div>Artificial intelligence has increasingly been used in interpreting medical images to support the timely treatment of diseases by providing early and accurate diagnosis. Pneumoconiosis is a tissue reaction that develops as a result of the accumulation of inorganic dust in the lungs. The most common types of pneumoconiosis include diseases such as coal worker's pneumoconiosis, silicosis, asbestosis, and siderosis. Silicosis, which has maintained its importance since the 1900s and has seen over 182,000 articles published in the last 10 years, is a global health problem. The automated detection and recognition of silicosis in lung computed tomography (CT) images can be considered the backbone of assisting the silicosis diagnosis process. Automated medical assistance systems developed using artificial intelligence can simplify the medical examination process and reduce the time required to start accurate treatment. Although the literature contains various studies that benefit silicosis diagnosis using chest X-ray images or pneumoconiosis diagnosis using CT images, there is not enough classification study that can particularly aid the diagnosis of silicosis in CT images.</div><div>The method of early detection of silicosis from chest radiographs and CT images has been a challenging task due to the high variability among pneumoconiosis readers. Based on the success of deep learning in the classification and segmentation of medical images, this study has shown that deep learning networks and transfer learning algorithms can detect silicosis with high accuracy by classifying CT images. The performance of the six algorithms examined in the study is compared, and the algorithm with the best performance is recommended. Performance criteria such as accuracy, precision, specificity, and F1-score of the algorithms used in the study were calculated. The accuracy rates of the models were obtained as 92.62 %, 93.03 %, 92.76 %, 95.38 %, 97.29 %, and 95.17 % for AlexNet, VGG16, ResNet50, InceptionV3, Xception, and DenseNet121, respectively. These results show that Xception outperformed the other algorithms and was the most successful algorithm in the automatic detection of silicosis with an accuracy rate of 97.29 %.</div><div>Additionally, a new dataset consisting of tomography images from silicosis patients is presented in this study. Experimental results have shown that transfer learning algorithms can significantly benefit the diagnosis of silicosis by successfully classifying CT images. The findings of the study highlight the clinical importance of artificial intelligence methods in medical image analysis and early disease diagnosis.</div></div>","PeriodicalId":72670,"journal":{"name":"Computer methods and programs in biomedicine update","volume":"6 ","pages":"Article 100166"},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142703968","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}