Yazar "Gok, Murat" seçeneğine göre listele

Listeleniyor 1 - 3 / 3
  • Küçük Resim Yok
    Öğe
    RESPONSE OF LYAPUNOV EXPONENTS TO DIFFUSION STATE OF BIOLOGICAL NETWORKS
    (Univ Zielona Gora Press, 2020) Altuntaş, Volkan; Gok, Murat; Kocal, Osman Hilmi
    The topologies of protein-protein interaction networks are uncertain and noisy. The network topology determines the reliability of computational knowledge acquired from noisy networks and can impose the deterministic and non-deterministic character of the resulting data. In this study, we analyze the effect of the network topology on Lyapunov exponents and its relationship with network stability. We define the methodology to convert the network data into signal data and obtain the Lyapunov exponents for a variety of networks. We then compare the Lyapunov exponent response and the stability results. Our technique can be applied to all types of network topologies as demonstrated with our experiments, conducted on both synthetic and real networks from public databases. For the first time, this article presents findings where Lyapunov exponents are evaluated under topological mutations and used for network analysis. Experimental results show that Lyapunov exponents have a strong correlation with network stability and both are correlatively affected by the network model. Hence we develop a novel coefficient, termed LEC, to measure the robustness of biological networks. LEC can be applied to real or synthetic biological networks rapidly. Results are a striking indication that the Lyapunov exponent is a potential candidate measure for network analysis.
  • Küçük Resim Yok
    Öğe
    Stability Analysis of Biological Networks' Diffusion State
    (Ieee Computer Soc, 2020) Altuntaş, Volkan; Gok, Murat; Kahveci, Tamer
    Computational knowledge acquired from noisy networks is not reliable and the network topology determines the reliability. Protein-protein interaction networks have uncertain topologies and noise that contain false positive and false negative edges at high rates. In this study, we analyze effects of the existing mutations in a network topology to the diffusion state of that network. To evaluate the sensitivity of the diffusion state, we derive the fitness measures based on the mathematically defined stability of a network. Searching for an influential set of edges in a network is a difficult problem. We handle the computational challenge by developing a novel metaheuristic optimization method and we find influential mutations time-efficiently. Our experiments, conducted on both synthetic and real networks from public databases, demonstrated that our method obtained better results than competitors for all types of network topologies. This is the first-time that the diffusion has been evaluated under topological mutations. Our analysis identifies significant biological results about the stability of biological - synthetic networks and diffusion state. In this manner, mutations in protein-protein interaction network topologies have a significant influence on the diffusion state of the network. Network stability is more affected by the network model than the network size.
  • Küçük Resim Yok
    Öğe
    The stability and fragility of biological networks: eukaryotic model organism Saccharomyces cerevisiae
    (Ieee, 2017) Altuntaş, Volkan; Gok, Murat
    Recent studies of biological networks show that these networks arc robust against the random or selective deletion of network nodes and I or edges. Ability to maintain performance of network under mutations is a key feature of live systems that has long been recognized. However, the molecular and cellular basis of this stability has just begun to be understood. Robustness is a key to understanding cellular complexity, illuminating design principles, and encouraging closer interaction between experiment and theory. A biological network mutation can be defined as the creation of a new network with k allowed network change operations for a given G network. While mutating the network, our goal is to observe the change in the measured distance estimate value after k changes of the defined distance measurement method M. In this study, the effects of edge deletion and edge insertion mutations on network topology and diffusion-based function estimation algorithms are investigated by using random mutation model on the protein-protein interaction network of eukaryote Saccharomyces cerevisiae yeast, containing 5936 nodes and 65139 edges. Experimental results shows that Saccharomyces cerevisiae protein-protein interaction network has high robustness against random mutations and that the generated mutations have no significant effect on network topology and estimation techniques.