Protein interaction networks and network stability analysis

Interactions of multiple molecules reveal biological activities. Biological networks provide great potential with unmatched knowledge to understand how basic processes control cellular activities. Computing information obtained from noisy networks is unreliable and depending on the type of information, network topology can have an effect. Protein-protein interaction (PPI) networks have indeterminate topologies and noise with high rates of false positive and false negative edges. In this thesis, it is suggested that possible mutations in a network topology may affect the information contained in this network and the algorithms running on the network. The aim of the thesis is to investigate how mutations in the network affect network reliability. In the thesis, in order to show the effect of mutations on both real and synthetic networks, a method has been developed to create highly effective artificial mutations in a particular network. Conformity measurements have been developed to evaluate the sensitivity of the target network to mutations. By comparing the fragility or stability of the networks, comparisons of synthetic networks and biological PPI networks against mutations have been performed. Relationships between Lypunov's exponentials and network reliability were investigated to see if there was a correlation between them. Lypunov's exponential-based attribute has been developed to ensure faster and easier detection of reliability results obtained by the method of generating a mutation in the network, which requires high processing load.