Research Outputs
Permanent URI for this communityhttps://hdl.handle.net/20.500.14288/2
Browse
12 results
Search Results
Publication Metadata only A kernel-based multilayer perceptron framework to identify pathways related to cancer stages(Springer Science and Business Media Deutschland GmbH, 2023) Mokhtaridoost, Milad; N/A; Department of Industrial Engineering; Soleimanpoor, Marzieh; Gönen, Mehmet; PhD Student; Faculty Member; Department of Industrial Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; 237468Standard machine learning algorithms have limited knowledge extraction capability in discriminating cancer stages based on genomic characterizations, due to the strongly correlated nature of high-dimensional genomic data. Moreover, activation of pathways plays a crucial role in the growth and progression of cancer from early-stage to late-stage. That is why we implemented a kernel-based neural network framework that integrates pathways and gene expression data using multiple kernels and discriminates early- and late-stages of cancers. Our goal is to identify the relevant molecular mechanisms of the biological processes which might be driving cancer progression. As the input of developed multilayer perceptron (MLP), we constructed kernel matrices on multiple views of expression profiles of primary tumors extracted from pathways. We used Hallmark and Pathway Interaction Database (PID) datasets to restrict the search area to interpretable solutions. We applied our algorithm to 12 cancer cohorts from the Cancer Genome Atlas (TCGA), including more than 5100 primary tumors. The results showed that our algorithm could extract meaningful and disease-specific mechanisms of cancers. We tested the predictive performance of our MLP algorithm and compared it against three existing classification algorithms, namely, random forests, support vector machines, and multiple kernel learning. Our MLP method obtained better or comparable predictive performance against these algorithms.Publication Metadata only Computational analysis of the binding free energy of H3K9ME3 peptide to the tandem tudor domains of JMJD2A(IEEE, 2010) N/A; Department of Chemical and Biological Engineering; Department of Computer Engineering; Department of Chemical and Biological Engineering; N/A; Keskin, Özlem; Gürsoy, Attila; Erman, Burak; Özboyacı, Musa; Faculty Member; Faculty Member; Faculty Member; PhD Student; Department of Computer Engineering; Department of Chemical and Biological Engineering; College of Engineering; College of Engineering; College of Engineering; Graduate School of Sciences and Engineering; 26605; 8745; 179997; N/AJMJD2A is a histone lysine demethylase enzyme which plays a prominent role in the development of prostate and esophageal squamous cancers. Consisting of a JmjC, a JmjN, two PHD and two tandem tudor domains, JMJD2A recognizes and binds to four different methylated histone peptides: H3K4me3, H4K20me3, H4K20me2 and H3K9me3, via its tudor domains. Of the four histone peptides, only recognition of the H3K4me3 and H4K20me3 by JMJD2A-tudor has been identified. In this study, we investigated the recognition of trimethylated H3K9 by the tandem tudor domains of JMJD2A. Using the molecular dynamics simulations, we performed normal mode and molecular mechanics - Poisson Boltzmann / generalized born - surface area (MM-PB/GB-SA) analysis to find the entropic and enthalpic contributions to binding free energy respectively. We showed that binding of the ligand is mainly driven by favorable van der Waals interactions made after complexation. Our findings suggest that, upon complex formation, H3K9me3 peptide adopts a similar binding mode and the same orientation with H3K4me3 peptide.Publication Metadata only Coupling between energy and residue position fluctuations in native proteins(IEEE, 2010) Department of Chemical and Biological Engineering; N/A; Erman, Burak; Gür, Mert; Faculty Member; PhD Student; Department of Chemical and Biological Engineering; College of Engineering; Graduate School of Sciences and Engineering; 179997; 216930The coupling between energy fluctuations and positional fluctuations in molecular dynamics trajectories of Crambin at 310 K is studied. Coupling with energy fluctuation is evaluated for both atomic positions and residue positions. Couplings show values which fluctuate around the previously proposed theoretical value under harmonic approximation. The magnitude of these correlations is in agreement, on the average, with the harmonic approximation. Additionally coupling between energy fluctuations and atom-atom distance fluctuations are evaluated. This coupling indicates how much each interaction among different atoms/residues is correlated with the protein's total energy fluctuations. Some atom's/residue's interactions have shown outstanding correlation. Moreover coupling of residue fluctuations between different modes is studied. © 2009 IEEE.Publication Metadata only Determination of the correspondence between mobility (rigidity) and conservation of the interface residues(IEEE, 2010) N/A; Department of Chemical and Biological Engineering; Department of Computer Engineering; N/A; Keskin, Özlem; Gürsoy, Attila; Makinacı, Gözde Kar; Faculty Member; Faculty Member; PhD Student; Department of Chemical and Biological Engineering; Department of Computer Engineering; College of Engineering; College of Engineering; Graduate School of Sciences and Engineering; 26605; 8745; N/AHot spots at protein interfaces may play specific functional roles and contribute to the stability of the protein complex. These residues are not homogeneously distributed along the protein interfaces; rather they are clustered within locally tightly packed regions forming a network of interactions among themselves. Here, we investigate the organization of computational hot spots at protein interfaces. A list of proteins whose free and bound forms exist is examined. Inter-residue distances of the interface residues are compared for both forms. Results reveal that there exist rigid block regions at protein interfaces. More interestingly, these regions correspond to computational hot regions. Hot spots can be determined with an average positive predictive value (PPV) of 0.73 and average sensitivity value of 0.70 for seven protein complexes.Publication Metadata only Inhibitor peptide design for NF- KB: Markov model and genetic algorithm(IEEE, 2010) N/A; Department of Computer Engineering; Department of Chemical and Biological Engineering; Ünal, Evrim Besray; Gürsoy, Attila; Erman, Burak; PhD Student; Faculty Member; Faculty Member; Department of Computer Engineering; Department of Chemical and Biological Engineering; Graduate School of Sciences and Engineering; College of Engineering; College of Engineering; N/A; 8745; 179997Two peptide design approaches are proposed to block activities of disease related proteins. First approach employs a probabilistic method; the problem is set as Markov chain. The possible binding site of target protein and a path on this binding site are determined. 20 natural amino acids and 400 dipeptides are docked to the selected path using the AutoDock software. The statistical weight matrices for the binding energies are derived from AutoDock results; matrices are used to determine top 100 peptide sequences with affinity to target protein. Second approach utilizes a heuristic method for peptide sequence determination; genetic algorithm (GA) with tournament selection. The amino acids are the genes; the peptide sequences are the chromosomes of GA. Initial random population of 100 chromosomes leads to determination of 100 possible binding peptides, after 8-10 generations of GA. Thermodynamic properties of the peptides are analyzed by a method that we proposed previously. NF-κB protein is selected as case-study.Publication Metadata only Interaction prediction of PDZ domains using a machine learning approach(IEEE, 2010) N/A; Department of Chemical and Biological Engineering; Department of Computer Engineering; N/A; Keskin, Özlem; Gürsoy, Attila; Kalyoncu, Sibel; Faculty Member; Faculty Member; Master Student; Department of Chemical and Biological Engineering; Department of Computer Engineering; College of Engineering; College of Engineering; Graduate School of Sciences and Engineering; 26605; 8745; N/AProtein interaction domains play crucial roles in many complex cellular pathways. PDZ domains are one of the most common protein interaction domains. Prediction of binding specificity of PDZ domains by a computational manner could eliminate unnecessary, time-consuming experiments. In this study, interactions of PDZ domains are predicted by using a machine learning approach in which only primary sequences of PDZ domains and peptides are used. In order to encode feature vectors for each interaction, trigram frequencies of primary sequences of PDZ domains and corresponding peptides are calculated. After construction of numerical interaction dataset, we compared different classifiers and ended up with Random Forest (RF) algorithm which gave the top performance. We obtained very high prediction accuracy (91.4%) for binary interaction prediction which outperforms all previous similar methods.Publication Metadata only Molecular communication transmitter architectures for the internet of bio-nano things(Institute of Electrical and Electronics Engineers Inc., 2022) Department of Electrical and Electronics Engineering; N/A; Akan, Özgür Barış; Civaş, Meltem; Faculty Member; PhD Student; Department of Electrical and Electronics Engineering; College of Engineering; Graduate School of Sciences and Engineering; 6647; N/AIn this study, we investigate the nanomaterial-based approach for developing practical molecular communication transmitters (MC-Txs) for the Internet of Bio-Nano Things (IoBNT) applications, which are expected to be unconventional in many aspects. In particular, we focus on the most pressing challenges for MC-Tx architectures, namely controlling information molecule release and molecule replenishment, together with other aspects, selective molecule release and molecule leakage mitigation. We discuss promising nanomaterials and also identify potential challenges and research directions.Publication Metadata only Nanosensor networks for smart health care(Elsevier, 2020) Abbasi, Naveed A.; Department of Electrical and Electronics Engineering; N/A; N/A; Department of Electrical and Electronics Engineering; Akan, Özgür Barış; Khan, Tooba; Civaş, Meltem; Çetinkaya, Oktay; Faculty Member; PhD Student; PhD Student; Other; Department of Electrical and Electronics Engineering; College of Engineering; Graduate School of Sciences and Engineering; Graduate School of Sciences and Engineering; College of Engineering; 6647; N/A; N/A; N/AAdvent of nanoscale sensors has paved the way for countless applications envisioned in the concept of a Smart City. In this chapter, we are focusing on one of the most fundamental requirements of the smart city, that is, smart health care. Great advancements in personal health care are expected with the emergence of nanosensing devices; however, single nanosensor is limited in its processing power and storage; thus we need to form network of nanosensors for any health-care application. In this chapter, we first elaborate the communication paradigms for nanosensor network. Moreover, we discuss various smart health-care applications such as smart drug delivery, body area network, implantable devices to treat injuries or malfunctions, and Internet of Nano Things. In the end, we highlight the implementation challenges for the nanosensor network for biomedical applications.Publication Metadata only Narrow escape problem in synaptic molecular communications(Association for Computing Machinery, Inc, 2022) Koca, Çağlar; Department of Electrical and Electronics Engineering; N/A; Akan, Özgür Barış; Civaş, Meltem; Faculty Member; PhD Student; Department of Electrical and Electronics Engineering; College of Engineering; Graduate School of Sciences and Engineering; 6647; N/AThe narrow escape problem (NEP) is a well-known problem with many applications in cellular biology. It is especially important to understand synaptic molecular communications. Active regions of synapses, also known as apposition zones, are connected to synaptic cleft through narrow slits, from which neurotransmitters can escape to or return from the cleft into the apposition zones. While neurotransmitters leakage into the cleft might be desired for the reuptake process, escaping neurotransmitters might trigger an undesired, i.e., false-positive or action potential in the post-synaptic terminal. Obtaining analytic solutions to NEPs is very challenging due to its geometry dependency. Slight alterations in either or both shape or the size of the hole and the outer volume may cause drastic changes in the solution. Thus, we need a simulation-based approach to solve NEPs. However, NEP also requires the size of the hole to be much smaller than the dimensions of the volume. Combined with the requirement for Brownian motion, where the step size is much smaller than the dimensions of the volume, simulations can be prohibitively long, even for modern computers. Therefore, in this work, we suggest a simulation algorithm that simultaneously satisfies the NEP and Brownian motion simulation requirements. Our simulation framework can be used to quantify the neurotransmitter leakage within synaptic clefts.Publication Open Access PrognosiT: pathway/gene set-based tumour volume prediction using multiple kernel learning(BioMed Central, 2021) Department of Industrial Engineering; N/A; Gönen, Mehmet; Bektaş, Ayyüce Begüm; Faculty Member; Department of Industrial Engineering; School of Medicine; College of Engineering; Graduate School of Sciences and Engineering; 237468; N/ABackground: identification of molecular mechanisms that determine tumour progression in cancer patients is a prerequisite for developing new disease treatment guidelines. Even though the predictive performance of current machine learning models is promising, extracting significant and meaningful knowledge from the data simultaneously during the learning process is a difficult task considering the high-dimensional and highly correlated nature of genomic datasets. Thus, there is a need for models that not only predict tumour volume from gene expression data of patients but also use prior information coming from pathway/gene sets during the learning process, to distinguish molecular mechanisms which play crucial role in tumour progression and therefore, disease prognosis. Results: in this study, instead of initially choosing several pathways/gene sets from an available set and training a model on this previously chosen subset of genomic features, we built a novel machine learning algorithm, PrognosiT, that accomplishes both tasks together. We tested our algorithm on thyroid carcinoma patients using gene expression profiles and cancer-specific pathways/gene sets. Predictive performance of our novel multiple kernel learning algorithm (PrognosiT) was comparable or even better than random forest (RF) and support vector regression (SVR). It is also notable that, to predict tumour volume, PrognosiT used gene expression features less than one-tenth of what RF and SVR algorithms used. Conclusions: PrognosiT was able to obtain comparable or even better predictive performance than SVR and RF. Moreover, we demonstrated that during the learning process, our algorithm managed to extract relevant and meaningful pathway/gene sets information related to the studied cancer type, which provides insights about its progression and aggressiveness. We also compared gene expressions of the selected genes by our algorithm in tumour and normal tissues, and we then discussed up- and down-regulated genes selected by our algorithm while learning, which could be beneficial for determining new biomarkers.