Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/395
Browse
14 results
Search Results
Article Predicting Respiratory Infection and Symptoms Development Using Gene Set Enrichment Scores and Machine Learning(Elsevier Sci Ltd, 2026) Aydin, Zafer; Isik, Yunus EmreRecent advancements in precision medicine enable personalized predictions grounded in individual-level genetic data. However, relying solely on a single type of data can decrease prediction accuracy and limit the biological interpretability of the resulting models. Incorporating predefined genetic knowledge, such as derived gene sets, can improve performance and provide deeper biological insights for complex diseases, including respiratory infections. This study aimed to evaluate the usability of enrichment scores (ES), calculated using gene sets from the Molecular Signatures Database (MSigDB), as a feature representation for machine learning models to predict respiratory viral infections and symptom development. In addition, the proposed feature representation approach was extensively compared with the de facto gene-level expression representation. A total of 36,834 predefined gene sets were compiled from the MSigDB, and their ES values were calculated. Experiments used the GSE73072 dataset from Gene Expression Omnibus, containing gene expression profiles before and after virus exposure. Various machine learning and feature selection algorithms were applied to ES-based and probe-level feature sets. The results showed that both feature representation approaches achieved an area under the precision-recall curve (AUPRC) value greater than 0.90 for all tasks. Compared with the Respiratory Viral DREAM Challenge leaderboard phase, our models showed a 14.8% improvement in pre-exposure predictions (T0) and a 17.4% improvement in symptom classification. Using enrichment scores as a feature representation generally resulted in better performance than probe-level representation when predicting respiratory infections and symptom development. Identifying key gene sets through feature selection and comparing them with essential genes for respiratory viruses enabled a more comprehensive analysis, providing deeper insights into the pathways that contribute to these predictions.Article Probing Adversarial Robustness of Protein Language Models: A Reproducible Case Study of ESM-2 Under Substitution-Based Attacks(AIUB Office of Research and Publication, 2026) Aydin, Zafer; Niloy, Md. Robiul Islam; Moazzam, Md. GolamArticle Frequency-Based Deep Occlusion Awareness Instance Segmentation(MDPI, 2026-02-26) Guzel, Yasin; Aydin, Zafer; Talu, Muhammed FatihOne major challenge faced by deep learning-based methods that detect target objects in the form of bounding boxes is object occlusion. High degrees of occlusion significantly diminish the accuracy of instance segmentation. Nonetheless, complex-valued Fourier descriptors can robustly represent object boundaries using minimal information. In this study, the impact of integrating Fourier descriptors-renowned for their strong representational capacity-with deep network models (UNet) that exhibit high generalization performance on instance segmentation accuracy was investigated. Within the scope of the research, nine network models were designed based on different strategies for utilizing frequency components. These variants fall into four strategy families: (i) UNet-style spectrum regression on fixed low-frequency windows (FUNet), (ii) magnitude-guided frequency selection/ROI construction (FUNet-Thr, FUNet-BBox), (iii) sequence models over tokenized FFT coefficients (BiLSTM Patch/Sorted), and (iv) encoder-only spectrum predictors with different depth/capacity (EncoderFFT1/2). To fairly evaluate the models' performance in segmenting objects subjected to disruptive factors (e.g., occlusion, blurring, noise), a specialized synthetic dataset was prepared. The task is formulated as single-target (single-instance), single-class segmentation. This dataset, automatically generated according to initial parameter values, contains images of objects moving at various speeds within a single frame. Among these models, the one termed FUNet, which relies on partial matching of central frequency components, achieved the highest segmentation accuracy despite the disruptive effects. Under the challenging Dataset 8 setting, the proposed FUNet achieved the highest overlap-based performance (Dice = 0.9329, IoU = 0.8842) among Attention U-Net, U-Net, and FourierNet, with statistically significant gains confirmed by paired per-image tests.Article Citation - WoS: 7Citation - Scopus: 8The Determination of Distinctive Single Nucleotide Polymorphism Sets for the Diagnosis of Behcet's Disease(IEEE Computer Soc, 2022-05-01) Isik, Yunus Emre; Gormez, Yasin; Aydin, Zafer; Bakir-Gungor, BurcuBehcet's Disease (BD) is a multi-system inflammatory disorder in which the etiology remains unclear. The most probable hypothesis is that genetic tendency and environmental factors play roles in the development of BD. In order to find the essential reasons, genetic changes on thousands of genes should be analyzed. Besides, there is a need for extra analysis to find out which genetic factor affects the disease. Machine learning approaches have high potential for extracting the knowledge from genomics and selecting the representative Single Nucleotide Polymorphisms (SNPs) as the most effective features for the clinical diagnosis process. In this study, we have attempted to identify representative SNPs using feature selection methods, incorporating biological information and aimed to develop a machine-learning model for diagnosing Behcet's disease. By combining biological information and machine learning classifiers, up to 99.64 percent accuracy of disease prediction is achieved using only 13,611 out of 311,459 SNPs. In addition, we revealed the SNPs that are most distinctive by performing repeated feature selection in cross-validation experiments.Article Citation - WoS: 5Citation - Scopus: 5Sample Reduction Strategies for Protein Secondary Structure Prediction(MDPI, 2019-10-18) Atasever, Sema; Aydin, Zafer; Erbay, Hasan; Sabzekar, MostafaPredicting the secondary structure from protein sequence plays a crucial role in estimating the 3D structure, which has applications in drug design and in understanding the function of proteins. As new genes and proteins are discovered, the large size of the protein databases and datasets that can be used for training prediction models grows considerably. A two-stage hybrid classifier, which employs dynamic Bayesian networks and a support vector machine (SVM) has been shown to provide state-of-the-art prediction accuracy for protein secondary structure prediction. However, SVM is not efficient for large datasets due to the quadratic optimization involved in model training. In this paper, two techniques are implemented on CB513 benchmark for reducing the number of samples in the train set of the SVM. The first method randomly selects a fraction of data samples from the train set using a stratified selection strategy. This approach can remove approximately 50% of the data samples from the train set and reduce the model training time by 73.38% on average without decreasing the prediction accuracy significantly. The second method clusters the data samples by a hierarchical clustering algorithm and replaces the train set samples with nearest neighbors of the cluster centers in order to improve the training time. To cluster the feature vectors, the hierarchical clustering method is implemented, for which the number of clusters and the number of nearest neighbors are optimized as hyper-parameters by computing the prediction accuracy on validation sets. It is found that clustering can reduce the size of the train set by 26% without reducing the prediction accuracy. Among the clustering techniques Ward's method provided the best accuracy on test data.Article Citation - Scopus: 6Network Intrusion Detection Based on Machine Learning Strategies: Performance Comparisons on Imbalanced Wired, Wireless, and Software-Defined Networking (SDN) Network Traffics(Turkiye Klinikleri, 2024-07-26) Hacilar, Hilal; Aydin, Zafer; Güngör, Vehbi ÇağrıThe rapid growth of computer networks emphasizes the urgency of addressing security issues. Organizations rely on network intrusion detection systems (NIDSs) to protect sensitive data from unauthorized access and theft. These systems analyze network traffic to detect suspicious activities, such as attempted breaches or cyberattacks. However, existing studies lack a thorough assessment of class imbalances and classification performance for different types of network intrusions: wired, wireless, and software-defined networking (SDN). This research aims to fill this gap by examining these networks’ imbalances, feature selection, and binary classification to enhance intrusion detection system efficiency. Various techniques such as SMOTE, ROS, ADASYN, and SMOTETomek are used to handle imbalanced datasets. Additionally, eXtreme Gradient Boosting (XGBoost) identifies key features, and an autoencoder (AE) assists in feature extraction for the classification task. The study evaluates datasets such as AWID, UNSW, and InSDN, yielding the best results with different numbers of selected features. Bayesian optimization fine-tunes parameters, and diverse machine learning algorithms (SVM, kNN, XGBoost, random forest, ensemble classifiers, and autoencoders) are employed. The optimal results, considering F1-measure, overall accuracy, detection rate, and false alarm rate, have been achieved for the UNSW-NB15, preprocessed AWID, and InSDN datasets, with values of [0.9356, 0.9289, 0.9328, 0.07597], [0.997, 0.9995, 0.9999, 0.0171], and [0.9998, 0.9996, 0.9998, 0.0012], respectively. These findings demonstrate that combining Bayesian optimization with oversampling techniques significantly enhances classification performance across wired, wireless, and SDN networks when compared to previous research conducted on these datasets. © 2024 Elsevier B.V., All rights reserved.Article Citation - WoS: 8Citation - Scopus: 10Lung Cancer Subtype Differentiation From Positron Emission Tomography Images(Tubitak Scientific & Technological Research Council Turkey, 2020-01-27) Ayyildiz, Oguzhan; Aydin, Zafer; Yilmaz, Bulent; Karacavus, Seyhan; Senkaya, Kubra; Icer, Semra; Kaya, Eser; Taşdemir, ArzuLung cancer is one of the deadly cancer types, and almost 85% of lung cancers are nonsmall cell lung cancer (NSCLC). In the present study we investigated classification and feature selection methods for the differentiation of two subtypes of NSCLC, namely adenocarcinoma (ADC) and squamous cell carcinoma (SqCC). The major advances in understanding the effects of therapy agents suggest that future targeted therapies will be increasingly subtype specific. We obtained positron emission tomography (PET) images of 93 patients with NSCLC, 39 of which had ADC while the rest had SqCC. Random walk segmentation was applied to delineate three-dimensional tumor volume, and 39 texture features were extracted to grade the tumor subtypes. We examined 11 classifiers with two different feature selection methods and the effect of normalization on accuracy. The classifiers we used were the k-nearest-neighbor, logistic regression, support vector machine, Bayesian network, decision tree, radial basis function network, random forest, AdaBoostM1, and three stacking methods. To evaluate the prediction accuracy we performed a leave-one-out cross-validation experiment on the dataset. We also considered optimizing certain hyperparameters of these models by performing 10-fold cross-validation separately on each training set. We found that the stacking ensemble classifier, which combines a decision tree, AdaBoostM1, and logistic regression methods by a metalearner, was the most accurate method for detecting subtypes of NSCLC, and normalization of feature sets improved the accuracy of the classification method.Article Citation - WoS: 10Citation - Scopus: 11Knowledge Based Response Correction Method for Design of Reconfigurable N-Shaped Microstrip Patch Antenna Using Inverse Anns(Wiley, 2015-12-18) Aoad, Ashrf; Simsek, Murat; Aydin, ZaferArtificial neural networks (ANNs) have been often used for engineering design problems. In this work, an inverse model of a reconfigurable N-shaped microstrip patch antenna which is formed by ANN is considered to find design parameters. For this task, knowledge-based response correction consists of two steps, which include generating response using multilayer perceptron as a first step and correcting this response using knowledge based methods such as source difference, prior knowledge input, and prior knowledge input with difference as a second step. The proposed antenna has four states of operation controlled by two Positive-Intrinsic-Negative (PIN) diodes with ON/OFF states. The two-step ANN models are inversely trained using the optimum of the resonant frequency parameter as the input and the physical dimensions of the proposed antenna as outputs of the multilayer perceptron. The outputs and, in some methods, the input parameters of the multilayer perceptron are sent as input to the knowledge-based models while the obtained outputs from the two steps are the results of the new physical dimensions of the redesigned reconfigurable antenna that will be compared and analyzed. This input/output complexity of the proposed reconfigurable antenna allows an accurate and fast inverse model to be developed with less training data. Users may use this antenna and its ANN models to develop new products in the market where any frequency in the operating region can be given to the input to result an appropriate form of the new reconfigurable antenna. Copyright (c) 2015 John Wiley & Sons, Ltd.Article Citation - WoS: 13Citation - Scopus: 21IGPRED: Combination of Convolutional Neural and Graph Convolutional Networks for Protein Secondary Structure Prediction(Wiley, 2021-05-25) Gormez, Yasin; Sabzekar, Mostafa; Aydin, ZaferThere is a close relationship between the tertiary structure and the function of a protein. One of the important steps to determine the tertiary structure is protein secondary structure prediction (PSSP). For this reason, predicting secondary structure with higher accuracy will give valuable information about the tertiary structure. Recently, deep learning techniques have obtained promising improvements in several machine learning applications including PSSP. In this article, a novel deep learning model, based on convolutional neural network and graph convolutional network is proposed. PSIBLAST PSSM, HHMAKE PSSM, physico-chemical properties of amino acids are combined with structural profiles to generate a rich feature set. Furthermore, the hyper-parameters of the proposed network are optimized using Bayesian optimization. The proposed model IGPRED obtained 89.19%, 86.34%, 87.87%, 85.76%, and 86.54% Q3 accuracies for CullPDB, EVAset, CASP10, CASP11, and CASP12 datasets, respectively.Article Citation - WoS: 4Citation - Scopus: 7IGPRED-Multitask: A Deep Learning Model to Predict Protein Secondary Structure, Torsion Angles and Solvent Accessibility(IEEE Computer Soc, 2023-03-01) Gormez, Yasin; Aydin, ZaferProtein secondary structure, solvent accessibility and torsion angle predictions are preliminary steps to predict 3D structure of a protein. Deep learning approaches have achieved significant improvements in predicting various features of protein structure. In this study, IGPRED-Multitask, a deep learning model with multi task learning architecture based on deep inception network, graph convolutional network and a bidirectional long short-term memory is proposed. Moreover, hyper-parameters of the model are fine-tuned using Bayesian optimization, which is faster and more effective than grid search. The same benchmark test data sets as in the OPUS-TASS paper including TEST2016, TEST2018, CASP12, CASP13, CASPFM, HARD68, CAMEO93, CAMEO93_HARD, as well as the train and validation sets, are used for fair comparison with the literature. Statistically significant improvements are observed in secondary structure prediction on 4 datasets, in phi angle prediction on 2 datasets and in psi angel prediction on 3 datasets compared to the state-of-the-art methods. For solvent accessibility prediction, TEST2016 and TEST2018 datasets are used only to assess the performance of the proposed model.