Rozprawy doktorskie na temat „Evolution on networks”

Evolutionary computation has been around ever since the late 50s. This thesis aims at elaborate on genetic algorithms, a subset of evolutionary computation, with particular regard to the field of neuroevolution, which is the application of GAs to the generation of functioning neural networks. The most widely adopted techniques are thereby explained and contrasted. The experimentation chapter finally shows an implementation of a genetic algorithm, inspired by existing algorithms, with the objective of optimizing a novel kind of artificial neural network.

A complex system is a system for which the statement "the whole is greater than the sum of its parts" holds. A network can be viewed as a backbone of a complex system. Combining the knowledge about the entities constituting the complex system with the properties of the interaction patterns we can get a better understanding of why the whole is greater than the sum. One of the purposes of network studies, is to relate the particular structural and dynamical properties of the network to the function it is designed to perform. In the present work I am briefly presenting some of the advances that have been achieved in the field of the complex networks together with the contributions which I have been involved in.

It is well known that the human brain is highly modular, having a structural and functional organization that allows the different regions of the brain to be reused for different cognitive processes. So far, this has not been fully addressed by artificial systems, and a better understanding of when and how modules emerge is required, with a broad framework indicating how modules could be reused within neural networks. This thesis provides a deep investigation of module formation, module communication (interaction) and module reuse during evolution for a variety of classification and prediction tasks. The evolutionary algorithm EPNet is used to deliver the evolution of artificial neural networks. In the first stage of this study, the EPNet algorithm is carefully studied to understand its basis and to ensure confidence in its behaviour. Thereafter, its input feature selection (required for module evolution) is optimized, showing the robustness of the improved algorithm compared with the fixed input case and previous publications. Then module emergence, communication and reuse are investigated with the modular EPNet (M-EPNet) algorithm, which uses the information provided by a modularity measure to implement new mutation operators that favour the evolution of modules, allowing a new perspective for analyzing modularity, module formation and module reuse during evolution. The results obtained extend those of previous work, indicating that pure-modular architectures may emerge at low connectivity values, where similar tasks may share (reuse) common neural elements creating compact representations, and that the more different two tasks are, the bigger the modularity obtained during evolution. Other results indicate that some neural structures may be reused when similar tasks are evolved, leading to module interaction during evolution.

Future generation cellular networks are expected to deliver an omnipresent broadband access network for an endlessly increasing number of subscribers. Long term Evolution (LTE) represents a significant milestone towards wireless networks known as 4G cellular networks. A key feature of LTE is the implementation of enhanced Radio Resource Management (RRM) mechanism to improve the system performance. The structure of LTE networks was simplified by diminishing the number of the nodes of the core network. Also, the design of the radio protocol architecture is quite unique. In order to achieve high data rate in LTE, 3rd Generation Partnership Project (3GPP) has selected Orthogonal Frequency Division Multiplexing (OFDM) as an appropriate scheme in terms of downlinks. However, the proper scheme for an uplink is the Single-Carrier Frequency Domain Multiple Access due to the peak-to-average-power-ratio (PAPR) constraint. LTE packet scheduling plays a primary role as part of RRM to improve the system’s data rate as well as supporting various QoS requirements of mobile services. The major function of the LTE packet scheduler is to assign Physical Resource Blocks (PRBs) to mobile User Equipment (UE). In our work, we formed a proposed packet scheduler algorithm. The proposed scheduler algorithm acts based on the number of UEs attached to the eNodeB. To evaluate the proposed scheduler algorithm, we assumed two different scenarios based on a number of UEs. When the number of UE is lower than the number of PRBs, the UEs with highest Channel Quality Indicator (CQI) will be assigned PRBs. Otherwise, the scheduler will assign PRBs based on a given proportional fairness metric. The eNodeB’s throughput is increased when the proposed algorithm was implemented.

The work presented in this Dissertation reflects a long-term human, professional and cultural path started some years ago when I first developed LogAnalysis, a tool for the analysis and visualization of criminal and social networks. Since then, I devoted myself to the development of frameworks, algorithms and techniques for supporting intelligence and law enforcement agencies in the task of unveiling the CN structure hidden in communication data, identifying the target offenders for their removal or selecting effective strategies to disrupt a criminal organization. In a natural way, I successively focused on the evaluation of the resilience of criminal networks and on the multiplex formalism, which takes into account the various relationships existing within a criminal organization. In this context I introduce criminal network analysis tools: LogAnalysis, LogViewer, Semantic viewer and Failure simulator. I have been involved in the design, modeling, and writing of all of the works presented. In particular, I have also developed and tested all the visual tools included therein. Finally, I introduce Multiplex PBFS (Mx-PBFS) a novel multi-threaded parallel Breadth-First Search algorithm for categorical and inter-layer couplings multiplex networks, and the framework CriMuxnet (still under development) for multilayer criminal networks analysis based on high-quality 3D visualizations of network data. CriMuxnet was designed to work in conjunction with a 3D computer graphics (CG) packages: Autodesk Maya or Blender. CriMuxnet exploits 3D engine features to significantly improve both exploratory search and visualization strategy.

This thesis work concerns about the Performance evolution of peer to peer networks, where we used different distribution technique’s of peer distribution like Weibull, Lognormal and Pareto distribution process. Then we used a network simulator to evaluate the performance of these three distribution techniques.During the last decade the Internet has expanded into a world-wide network connecting millions of hosts and users and providing services for everyone. Many emerging applications are bandwidth-intensive in their nature; the size of downloaded files including music and videos can be huge, from ten megabits to many gigabits. The efficient use of network resources is thus crucial for the survivability of the Internet. Traffic engineering (TE) covers a range of mechanisms for optimizing operational networks from the traffic perspective. The time scale in traffic engineering varies from the short-term network control to network planning over a longer time period.Here in this thesis work we considered the peer distribution technique in-order to minimise the peer arrival and service process with three different techniques, where we calculated the congestion parameters like blocking time for each peer before entering into the service process, waiting time for a peers while the other peer has been served in the service block and the delay time for each peer. Then calculated the average of each process and graphs have been plotted using Matlab to analyse the results

The aim of this master thesis is to study algorithms for automatically tuning antenna parameters to improve the performance of the radio access part of a telecommunication network and user experience. There are four dierent optimization algorithms, Stepwise Minimization Algorithm, Random Search Algorithm, Modied Steepest Descent Algorithm and Multi-Objective Genetic Algorithm to be applied to a model of a radio access network. The performances of all algorithms will be evaluated in this thesis. Moreover, a graphical user interface which is developed to facilitate the antenna tuning simulations will also be presented in the appendix of the report.

Biochemical regulatory networks are involved in many essential processes of life, such as maintaining homeostatic conditions, cell-cycle division, and responding to environmental stimuli. Such networks contain a staggering degree of complexity, inhibiting our ability to understand how they function. Two major challenges in quantitative biology are addressed in this thesis. First, the problem of identifying regulatory models from observational data. Second, characterizing how regulatory networks evolve. Making use of a simple regulatory model that is based on piecewise linear differential equations, investigations into these questions are undertaken. A novel method for inferring regulatory networks from knowledge of the dynamics is presented. This result is used to derive an atlas of networks with highly robust limit cycle dynamics in dimensions 3,4 and 5. Additionally, a new theoretical approach is used to identify two families of regulatory networks: cyclic, negative feedback and sequential disinhibition, with robust periodic dynamics that exist in all higher dimensions too. Smooth generalizations of the piecewise linear case, called continuous homologues, are also considered. For the family of cyclic, negative feedback networks it is shown that for the continuous homologue, a limit cycle exists and provided it is hyperbolic, it is asymptotically stable. Regulatory networks share many features across different species, raising the question of how they evolved. A simple evolutionary model in which piecewise linear networks are subject to mutation processes with selective pressure for chaotic dynamics is considered. Investigation into the evolutionary process reveals that the robustness (insensitivity of dynamics to mutations of the components of the network) has an important impact on the ability to innovate increasingly chaotic dynamics. An explicit analytical description of how evolutionary processes
Biochemical réglementaire des réseaux sont impliqués dans de nombreux processusessentiels de la vie, tels que les conditions en restant homéostatique, le cycle dedivision cellulaire, et de répondre 'a l'environnement stimuli. Ces réseaux contiennentun grand degré de complexité, qui nous empéche de comprendre comment ils fonctionnent.Deux défis majeurs dans la biologie quantitative sont abordées dans cetteth'ese. Tout d'abord, le probl'eme de l'identification des mod'eles de réglementation 'apartir de données d'observation. Deuxi'emement, révélant les processus par lesquelsles réseaux de régulation évoluent. L'utilisation d'une simple réglementation mod'elequi est basé sur les équations différentielles linéaires par morceaux; les enquêtes surces questions sont mises en oeuvre.Une nouvelle méthode pour déduire des réseaux de réglementation de la connaissancede la dynamique est présenté. Ce résultat est utilisé pour tirer un atlasdes réseaux tr'es robuste cycle de limite dynamiques dans les dimensions 3,4 et 5.En outre, une nouvelle approche théorique est également utilisé pour identifier deuxfamilles de réseaux de régulation: la rétroaction négative de réseaux et séquentieldisinhibition, avec de solides périodiques dynamiques qui existe aussi dans toutes lesdimensions supérieures aussi. Les généralisations de linéaire par lisse morceaux del'esp'ece, appelée continue homologues, sont également considérée. Pour la famillede cyclique, la rétroaction négative de réseaux, il est montré que, dans le homologuecontinue, un cycle de limite existe et il est prévu hyperbolique, il est alors asymptotiquementstable.Réglementaire des réseaux part de nombreuses fonctionnalités 'a travers les différentesesp'eces, ce qui soul'eve la question de savoir comment ils évoluent. Un mod'ele simpleévolutif dans lequel les réseaux sont linéaires par morceaux soumis 'a$

This thesis describes a novel approach to the evolution of Modular Artificial Neural Networks. Standard Evolutionary Algorithms, used in this application include: Genetic Algorithms, Evolutionary Strategies, Evolutionary Programming and Genetic Programming; however, these often fail in the evolution of complex systems, particularly when such systems involve multi-domain sensory information which interacts in complex ways with system outputs. The aim in this work is to produce an evolutionary method that allows the structure of the network to evolve from simple to complex as it interacts with a dynamic environment. This new algorithm is therefore based on Incremental Evolution. A simulated model of a legged robot was used as a test-bed for the approach. The algorithm starts with a simple robotic body plan. This then grows incrementally in complexity along with its controlling neural network and the environment it reacts with. The network grows by adding modules to its structure - so the technique may also be termed a Growth Algorithm. Experiments are presented showing the successful evolution of multi-legged gaits and a simple vision system. These are then integrated together to form a complete robotic system. The possibility of the evolution of complex systems is one advantage of the algorithm and it is argued that it represents a possible path towards more advanced artificial intelligence. Applications in Electronics, Computer Science, Mechanical Engineering and Aerospace are also discussed.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Physics, 2005.
Includes bibliographical references (p. 57-58).
In this thesis, I study the process of evolution of the gene regulatory network in Escherichia coli. First, I characterize the portion of the network that has been documented, and then I simulate growth of the network. In this study, I assume that the network evolves by gene duplication and divergence. Initially, the duplicated gene will retain its old interactions. As the gene accumulates mutations, it gains new interactions and may or may not lose the old interactions. I investigate evidence for the duplication-divergence model by looking at the homology and regulatory networks in E. coli and propose a simple duplication-divergence model for growth. The results show that this simple model cannot fully account for the complexity in the real network fragment as measured by conventional metrics.
by Juhi Kiran Chandalia
S.M.

In this work, the evolution of sex determination gene networks is inves tigated using a modelling approach. Recent evidence indicates that an in crease in the complexity of interactions has played an important role in gene network evolution. Sex determination mechanisms offer a good model for studying gene network evolution because, among other reasons, they evolve rapidly. In chapter 2, the potential for evolutionary change of the existing Drosophila sex determination gene network is considered. With the aid of a synchronous logical model, theoretical concepts such as a network-specific form of mutation are defined, as well as a notion of functional equivalence between networks. Applying this theoretical framework to the sex deter mination mechanism, it is found that sex determination networks generally exist within large sets of functionally equivalent networks all of which satisfy the sex determination task. These large sets are in turn composed of sub sets which are mutationally related, suggesting a high degree of flexibility is available without compromising the core functionality. The technique for finding functional equivalence between networks suggests a general method for gene network reconstruction, which is explored in chapter 3. Lastly, in chapters 4 and 5, a hierarchical model is presented which integrates popu lation genetics techniques with network dynamics. This model consists of a core population genetics simulation within which parameters such as the sex and fitness of the genotype are calculated from the corresponding network dynamics. The model is used to investigate the early evolution of sex deter mination networks. Following from a hypothesis proposed by Wilkins (1995), the assumption is made that sex determination networks have evolved in a retrograde manner from bottom to top. Starting from the simplest possible ancestral system, based on a single locus, we explore the way in which more complex systems, involving two or three loci, could have evolved.

In this thesis we propose some new approaches to the study of complex networks, and apply them to multiple domains, focusing in particular on protein-protein interaction networks. We begin by examining the roles of individual proteins; specifically, the influential idea of 'date' and 'party' hubs. It was proposed that party hubs are local coordinators whereas date hubs are global connectors. We show that the observations underlying this proposal appear to have been largely illusory, and that topological properties of hubs do not in general correlate with interactor co-expression, thus undermining the primary basis for the categorisation. However, we find significant correlations between interaction centrality and the functional similarity of the interacting proteins, indicating that it might be useful to conceive of roles for protein-protein interactions, as opposed to individual proteins. The observation that examining just one or a few network properties can be misleading motivates us to attempt to develop a more holistic methodology for network investigation. A wide variety of diagnostics of network structure exist, but studies typically employ only small, largely arbitrarily selected subsets of these. Here we simultaneously investigate many networks using many diagnostics in a data-driven fashion, and demonstrate how this approach serves to organise both networks and diagnostics, as well as to relate network structure to functionally relevant characteristics in a variety of settings. These include finding fast estimators for the solution of hard graph problems, discovering evolutionarily significant aspects of metabolic networks, detecting structural constraints on particular network types, and constructing summary statistics for efficient model-fitting to networks. We use the last mentioned to suggest that duplication-divergence is a feasible mechanism for protein-protein interaction evolution, and that interactions may rewire faster in yeast than in larger genomes like human and fruit fly. Our results help to illuminate protein-protein interaction networks in multiple ways, as well as providing some insight into structure-function relationships in other types of networks. We believe the methodology outlined here can serve as a general-purpose, data-driven approach to aid in the understanding of networked systems.

Under urbanisation, transport infrastructures may be improved when urban population grows. Meanwhile, land use patterns may vary and this urban dynamics may drive variations in mode choice of commuters and spatial features of transport networks. Empirical studies have observed scaling laws between the amount of transport infrastructures and city sizes. This thesis is aiming to provide a modelling framework for the analytical investigation of network growth and present some empirical observations of the variation in spatial features of transport networks. First, a simple linear monocentric city model is formulated and the global performance of transport systems is derived. Two cases according to strategies of urban intensification and sprawl have been studied to examine the consequence of the scaling-law growth in transport infrastructures. Second, this thesis proposes a modelling framework. The framework includes two congestible modes, the scaling-law growth of transport infrastructures and housing allocation of residents so that phenomena under urban dynamics could be modelled. The experiments show that the proposed modelling framework could investigate the trade-off of investment on the highway and public transport system. Third, empirical observations of spatial features in transport networks are reported in this thesis. The thesis measures circuity of transport networks, because this indicator could examine how aggregate transport networks are and the efficiency of network structures. Then research methods that can deal with several data sources are developed. The empirical observation shows that there is an exponential decay between the circuity and travel time in public transport networks. Meanwhile, this thesis also presents that the average circuity in road networks is less than that in public transport networks for the same sample of trips, which to some extent show the difference of spatial features between road and public transport networks. Additionally, correlations between circuity, accessibility and mode share are analysed.

The events of September 11,2001 followed by the oil price hike and the economic crisis of 2008, have lead to a drop in the demand for air travel. Airlines have attempted to return to profitability by cutting service in certain unattractive routes and airports. Simultaneously, delays and excess demand at a few major hubs have lead to airline introducing service at reliever airports. This dissertation attempts to capture the changes in the airline network by utilizing a supply-demand framework.
Ph. D.

It has been argued that much of evolution takes place in the absence of fitness gradients. Such periods of evolution can be analysed by examining the mutational network formed by sequences of equal fitness, that is, the neutral network. It has been demonstrated that, in large populations under a high mutation rate, the population distribution over the neutral network and average mutational robustness are given by the principal eigenvector and eigen- value, respectively, of the network's adjacency matrix. However, little progress has been made towards understanding the manner in which the topology of the neutral network influences the resulting population distribution and robustness. In this work, we build on recent results from spectral graph theory and utilize numerical methods to enhance our understanding of how populations distribute themselves over neutral networks. We demonstrate that, in the presence of certain topological features, the population will undergo an exploration catastrophe and become confined to a small portion of the network. We further derive approximations, in terms of mutational biases, for the population distribution and average robustness in networks with a homogeneous structure. The applicability of these results is explored, first, by a detailed review of the literature in both evolutionary computing and biology concerning the structure of neutral networks. This is extended by studying the actual and predicted population distribution over the neutral networks of H1N1 and H3N2 influenza haemagglutinin during seasons between 2005 and 2016. It is shown that, in some instances, these populations experience an exploration catastrophe. These results provide insight into the behaviour of populations on neutral networks, demonstrating that neutrality does not necessarily lead to an exploration of genotype/phenotype space or an associated increase in population diversity. Moreover, they provide a plausible explanation for conflicting results concerning the relationship between robustness and evolvability.

As a result of the constant efforts to improve mobile system performance and spectral efficiency, the 3GPP standardization forum is currently defining new architectural and functional requirements that hope to ensure long-term evolution (specifically defined as the “Long-Term Evolution (LTE) concept”) and general future competitiveness of the 2G and 3G radio access technologies. Previous discussions on LTE efficiency have been focused on general assumptions on signaling overhead and overall system capacity, based on experience from existing mobile systems. However, as 3GPP standardization has become more mature (although not yet settled), there is a need to investigate how different potential LTE services will be affected by the use of available overhead information and basic scheduling algorithms. This thesis investigates the lower protocol layers’ overhead impacts on the downlink for different packet switched services, in an LTE radio access network (RAN). Results show that the use of RTP/TCP/IP header compression (ROHC) is the single most important factor to reduce payload overhead, for packet sizes of ~1kB or smaller. However, for packets larger than ~1 kB, the use of ROHC becomes insignificant. Protocol headers – including the AMR frame header, RLC/MAC headers, and CRC where applicable – remain the largest part of payload overhead regardless of packet size and header compression (ROHC). For VoIP over the UDP protocol (with ROHC), RLC/MAC headers constitute the largest part of protocol headers. For TCP/IP applications (without ROHC), TCP/IP headers are predominant. Services that require packet sizes beyond ~1 kB will require about the same power per payload bit regardless of percentage of payload overhead.
Som ett resultat av ständiga ansträngningar att förbättra såväl prestanda som spektrumeffektivitet för mobila system, definierar 3GPPs standardiseringsforum nya krav på arkitektur och funktionalitet. Dessa är avsedda att säkerställa långsiktig utveckling (explicit definierat som konceptet “Long-Term Evolution (LTE)”, samt framtida konkurrenskraft för både 2G och 3G som radioaccess-teknologier. Tidigare diskussioner rörande effektivitet inom LTE har fokuserat på allmänna antaganden vad gäller kontrolldata för signallering och övergripande systemprestanda. Dessa har i sin tur baserats på erfarenheter från existerande mobilsystem. När standardiseringen inom 3GPP mognar uppstår nu ett behov av att undersöka hur olika tjänster inom LTE påverkas, av såväl hur man använder den kontrollinformation som finns tillgänglig, som av basala algoritmer for schemaläggning av resurser. Denna rapport undersöker påverkan från lägre protokoll-lagers kontrollinformation på nerlänken hos olika paket-kopplade tjänster inom ett radioaccessnät för LTE. Resultaten visar att användandet av ROHC (som packar kontrollinformation för protokollen RTP/TCP/IP), är det ensamt viktigaste bidraget till minskad kontrollinformation i relation till informationsbitar för paketstorlekar upp till c:a 1kB. För större paket är vinsten med ROHC dock försumbar. Kontrollinformation för protokoll – inkluderat data avsett för AMR-tal-ramen, RLC/MAC-protokollen, samt CRC – utgör för övrigt en stor del av kontrollinformationen relativt informationsbitar, oavsett paketstorlek och packning av kontrolldata. Tjänster som kräver paketstorlekar på över c:a 1 kB kräver uppskattningsvis samma mängd energi per informationsbit, oavsett andelen kontrollinformation.

In this thesis, we address problems encountered in complex network analysis using graph theoretic methods. The thesis specifically centers on the challenge of how to characterize the structural properties and time evolution of graphs. We commence by providing a brief roadmap for our research in Chapter 1, followed by a review of the relevant research literature in Chapter 2. The remainder of the thesis is structured as follows. In Chapter 3, we focus on the graph entropic characterizations and explore whether the von Neumann entropy recently defined only on undirected graphs, can be extended to the domain of directed graphs. The substantial contribution involves a simplified form of the entropy which can be expressed in terms of simple graph statistics, such as graph size and vertex in-degree and out-degree. Chapter 4 further investigates the uses and applications of the von Neumann entropy in order to solve a number of network analysis and machine learning problems. The contribution in this chapter includes an entropic edge assortativity measure and an entropic graph embedding method, which are developed for both undirected and directed graphs. The next part of the thesis analyzes the time-evolving complex networks using physical and information theoretic approaches. In particular, Chapter 5 provides a thermodynamic framework for handling dynamic graphs using ideas from algebraic graph theory and statistical mechanics. This allows us to derive expressions for a number of thermodynamic functions, including energy, entropy and temperature, which are shown to be efficient in identifying abrupt structural changes and phase transitions in real-world dynamical systems. Chapter 6 develops a novel method for constructing a generative model to analyze the structure of labeled data, which provides a number of novel directions to the study of graph time-series. Finally, in Chapter 7, we provide concluding remarks and discuss the limitations of our methodologies, and point out possible future research directions.

I present a method for inferring a protein interdependency network, based on correlated evolution between proteins on a large phylogeny of 72 diverse eukaryotic species. My original contribution is in the span of the phylogenetic tree used to generate the network: similar studies have concentrated on more localised regions in the tree of life and have been undertaken with more limited intent. I show that the whole-eukaryotic correlated evolution network is a real network and has interesting features of its own. The method can be broken down into three major, sequential parts: binary trait derivation, phylogenetic inference and likelihood analysis. I describe the implementation of a reciprocal BLAST protocol for the inference of a binary trait matrix corresponding to the presence or absence of orthologues in each species in the analysis, based on a reference human proteome. Rows in the matrix correspond to the reference proteins, columns to species: entries are 1, denoting presence, or 0, denoting absence. The matrix that I derive is mapped onto a phylogeny of the same set of species, to facilitate the detection of correlated evolution between proteins, or orthologous sets thereof, based on the pattern of gains and losses. The phylogeny is inferred from a set of genes, which are selected according to the criterion that they be present in all 72 species. 15 genes meeting this criterion are identified from the trait matrix; 14 of them are aligned and used for phylogenetic inference. The inference itself is performed using the program BayesPhylogenies, which implements a phylogenetic mixture model using a Markov Chain Monte Carlo (MCMC) method. A consensus phylogeny (tree) is calculated after the chain has been run for many millions of iterations; trees based on the genes individually were also inferred, for purposes of comparison. I use the program BayesTraits to perform a likelihood analysis on pairs of proteins from the trait matrix. This method detects correlated evolution by means of a likelihood ratio statistic, relating the likelihood of the two proteins having evolved independently, to the likelihood of their having evolved in an interdependent, or correlated, fashion. If the likelihood ratio statistic exceeds a certain threshold, this is interpreted as the signature of correlated evolution. Using presumptively interacting protein pairs from the Human Protein Reference Database, and a control (or null) set of pairs where no interaction is expected, I present evidence for the efficacy of the method in detecting correlated evolution. I proceed to infer a network based on correlated evolution, wherein each link represents an instance of pairwise correlation, and demonstrate that a power law gives a good fit to the distribution of nodal degree within the network, which is also the case for a network of presumptive protein interactions with no filter for correlated evolution. Finally, I infer a new equation to characterise the evolutionary rules which fashioned the network. I propose a method for testing the equation, and discuss future directions.

Numerous studies have shown that interpersonal communication acts as animportant channel for gathering information. But if we wish to rely oninterpersonal communication, we still need to figure out how to determinethe right person to ask. Usually we cannot find the potential expert(s)directly, and we need some assistance from our friends or friends'friends to locate them. The phenomenon of Six Degrees of Separationindicates that it is possible to use some intelligent software agents,who can interpret the links between people and follow only therelevant one, to find the desired experts quickly. A computational model of agent-based referral networks was proposed to assistand simplify the users to find potential experts for a specified topic in aperson-to-person social network, in which each user is assigned a softwareagent, and software agents help automate the process by a series of ``referralchains''. Unlike most previous approaches, our architecture is fullydistributed and includes agents who preserve the privacy and autonomy oftheir users. These agents learn models of each other in terms of expertise(ability to produce correct domain answers), and sociability (ability toproduce accurate referrals). We study this framework experimentally to seethe effects that the different variables have on each other and the efficiencyof the referral networks. Furthermore, a social mechanism of reputation management was proposed tohelp agents (users) avoid interaction with undesirable participants inthe referral networks. The mathematical theory of evidence is used torepresent and propagate the reputation information in a referral network.Our approach adjusts the ratings of agents based on their observations aswell the testimony from others. Moreover, we conducted several experimentsto study the reputation management in different settings. Social mechanismsare even more important when some centralized reputation managementmechanisms, i.e., trusted third parties, are not available. Our specificapproach to reputation management leads to a decentralized society in whichagents help each other weed out undesirable players.

In contrast to protein structure and to metabolic networks, an organism's transcriptional regulatory network (TRN) is highly plastic and has been reported to be the target of numerous mutations affecting the phenotype. The ascomycetes fungi are an ideal phylogenetic branch to look at TRN evolution by assessing the conservation of transcription factors (TF) and TF binding sites. In this thesis I 1) describe new tools to assess protein function in Candida albicans, 2) report a novel mode of TRN evolution that we named TF substitution that occurred in the evolution of the ribosomal protein (RP) regulon and 3) explore the large-scale transcription network reorganization of two RP TRNs with distinct wirings and TF substitutions. This work leads to a broader understanding of the span of TRN evolution at the organismal level in ascomycetes. These findings illustrate the extent of TRN plasticity in the evolution of fungi. They also raise questions regarding the rate and importance of TRN changes in the evolution of eukaryotes.
Contrairement à la structure des domaines protéiques et aux réseaux métaboliques qui sont difficilement interchangeables, le réseau de régulation transcriptionnelle (TRN) d'un organisme se doit d'être extrêmement plastique et peut être sujet à plusieurs types de mutations influant sur le phénotype. Les levures ascomycètes constituent un groupe phylétique idéal pour évaluer la conservation des facteurs de transcription (TFs) et de leurs sites de liaison sur l'ADN. Dans cette thèse, 1) je décris de nouveaux outils permettant l'analyse fonctionnelle des protéines de C. albicans, 2) Je rapporte le premier exemple d'un changement dans le TRN baptisé substitution de TF qui s'est produit dans l'évolution du régulon des protéines ribosomales (RP) et 3) j'explore les modalités de réorganisation du réseau transcriptionnel dans un système recâblé suite à une substitution de TF. Ces travaux pavent la voie à une étude élargie de l'évolution du TRN chez les ascomycètes. Ils illustrent aussi l'extrême plasticité du TRN ribosomal dans l'évolution des ascomycètes et soulèvent d'intéressantes questions quant à l'importance des changements du réseau régulatoire dans l'évolution des eucaryotes.

Language is in constant flux, whether through the creation of new terms or the changing meanings of existing words. The process by which language change happens is through complex reinforcing interactions between individuals and the social structures in which they exist. There has been much research into language change and evolution, though this has involved manual processes that are both time consuming and costly. However, with the growth in popularity of osn, for the first time, researchers have access to fine-grained records of language and user interactions that not only contain data on the creation of these language innovations but also reveal the inter-user and inter-community dynamics that influence their adoptions and rejections. Having access to these osn datasets means that language change and evolution can now be assessed and modelled through the application of computational and machine-learning-based methods. Therefore, this thesis looks at how one can detect and predict language change in osn, as well as the factors that language change depends on. The answer to this over-arching question lies in three core components: first, detecting the innovations; second, modelling the individual user adoption process; and third, looking at the collective adoption across a network of individuals. In the first question, we operationalise traditional language acceptance heuristics (used to detect the emergence of new words) into three classes of computation time-series measures computing the variation in frequency, form and/or meaning. The grounded methods are applied to two osn, with results demonstrating the ability to detect language change across both networks. By additionally applying the methods to communities within each network, e.g. geographical regions, on Twitter and Subreddits in Reddit, the results indicate that language variation and change can be dependent on the community memberships. The second question in this thesis focuses on the process of users adopting language innovations in relation to other users with whom they are in contact. By modelling influence between users as a function of past innovation cascades, we compute a global activation threshold at which users adopt new terms dependent on exposure to them from their neighbours. Additionally, by testing the user interaction networks through random shuffles, we show that the time at which a user adopts a term is dependent on the local structure; however, a large part of the influence comes from sources external to the observed osn. The final question looks at how the speakers of a language are embedded in social networks, and how the networks' resulting structures and dynamics influence language usage and adoption patterns. We show that language innovations diffuse across a network in a predictable manner, which can be modelled using structural, grammatical and temporal measures, using a logistic regression model to predict the vitality of the diffusion. With regard to network structure, we show how innovations that manifest across structural holes and weak ties diffuse deeper across the given network. Beyond network influence, our results demonstrate that the grammatical context through which innovations emerge also play an essential role in diffusion dynamics - this indicates that the adoption of new words is enabled by a complex interplay of both network and linguistic factors. The three questions are used to answer the over-arching question, showing that one can, indeed, model language change and forecast user and community adoption of language innovations. Additionally, we also show the ability to apply grounded models and methods and apply them within a scalable computational framework. However, it is a challenging process that is heavily influenced by the underlying processes that are not recorded within the data from the osns.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2008.
Includes bibliographical references (p. 129).
We study three types of communication data--emails, phone calls, and meetings-in a modern IT firm. Using network descriptive statistics, we show how communication networks in an organization differ from random networks and other social networks. We also compare and contrast the three types of communication networks. Using Quadratic Assignment Procedure (QAP), Multiple Regression Quadratic Assignment Procedure (MRQAP) and Exponential Random Graph Models (ERGM), we identify significant factors affecting the size and shape of communication networks. These parameters include organizational structure, homophily, job position, and physical proximity. We record the evolution of the networks and discuss how the factors affecting initial network growth differ from the steady state of the network.Erik Brynjolfsson George and Sandi Schussel Professor of Management Director, Center for Digital Business, MIT Sloan School of Management
by Jichao Qian.
M.Eng.

Various phenomena within the information systems discipline can be studied using the social network paradigm that views social entities as nodes with links between them. The social network analysis (SNA) theory has applications in knowledge management, computer mediated communications, security informatics, and other domains. Challenges in SNA can be classified into three broad areas: instance modeling, evolution modeling, and predictive modeling. Instance modeling focuses on the study of static network properties, evolution modeling examines factors behind network growth, and predictive modeling is concerned with identification of hidden and future network links. This dissertation presents four essays that address these challenges with empirical studies in knowledge management and security informatics.The first essay on instance and evolution modeling contributes to SNA theory by examining a real-world network that contains interactions between thirty thousand individuals. The study is among the few that empirically examine large human-only networks and verify the presence of small-world properties and scale-free distributions. In addition, it proposes a novel application of a network evolution model to examine the growth of networks across geographical boundaries.The second essay on evolution modeling proposes a methodology to identify significant link-formation facilitators. The study found that homophily in age, gender, and race were not significant factors in predicting future links between individuals in dark networks. These results contradicted some previous studies in the same domain that used smaller datasets to study the phenomena.The third essay focuses on evolution and predictive modeling and examines the role of inventor status on the selection of knowledge recombined to produce innovation. A new network measure based on random walks and team identification (RWT) is proposed to model knowledge flow. It is found that inventor status as measured by RWT has a positive relationship with the likelihood of a future citation link to the inventor.The fourth essay focuses on predictive modeling. A modified mutual information formulation is proposed to identify hidden links between nodes based on heuristics of time and location of previous co-occurrences. An evaluation of the proposed technique showed that it performed better in predicting hidden links than other co-occurrence based methods.

Recent studies on S. cerevisiae showed that essential genes, genes coding for protein complexes and network hubs are singleton, i.e. their duplications are negatively selected. These genes are intrinsically fragile toward perturbations, since dosage modifications yield to alterations in protein function, causing phenotypic aberrations that may affect the whole cell function. Unlike yeast, mammalian duplicated genes mostly encode highly connected proteins, while essentiality is not correlated with duplicability. This difference suggests that dosage-sensitive genes could duplicate at a certain point in evolution, likely favoring the progressive increase in genomic complexity. To understand when and how gene duplicability changed in evolution, we compared gene and network properties in several species from bacteria to primates. The origin and conservation of a gene significantly correlates with the properties of the encoded protein in the protein interaction network. All networks preserve a core of singleton and central hubs that originated early in evolution, are highly conserved, and accomplish basic biological functions. Another group of hubs appeared in metazoans and duplicated in vertebrates, mostly through vertebrate-specific whole genome duplication. Such hubs are frequently target of microRNAs and show tissue-selective expression, suggesting that these are alternative mechanisms to control their dosage. Our study shows how networks modified during evolution and contributes to explain the occurrence of somatic genetic diseases, such as cancer, in terms of network perturbations. Determining the evolutionary characteristics of cancer genes and their position inside the protein interaction network helps understanding the importance of these properties in tumorigenesis.

The World Wide Web (or Web for short) is the largest client-server computing system commonly available, which is used through its widely accepted universal client (the Web browser) that uses a standard communication protocol known as the HyperText Transfer Protocol (HTTP) to display information described in the HyperText Markup Language (HTML). The current Web computing model allows the execution of server-side applications such as Servlets and client-side applications such as Applets. However, it offers limited support for another model of network computing where users would be able to use remote, and perhaps more powerful, machines for their computing needs. The client-server model enables anyone with a Web-enabled device ranging from desktop computers to cellular telephones, to retrieve information from the Web. In today's information society, however, users are overwhelmed by the information with which they are confronted on a daily basis. For subscribers of mobile wireless data services, this may present a problem. Wireless handheld devices, such as cellular telephones are connected via wireless networks that suffer from low bandwidth and have a greater tendency for network errors. In addition, wireless connections can be lost or degraded by mobility. Therefore, there a need for entities that act on behalf of users to simplify the tasks of discovering and managing network computing resources. It has been said that software agents are a solution in search of a problem. Mobile agents, however, are inherently distributed in nature, and therefore they represent a natural view of a distributed system. They provide an ideal mechanism for implementing complex systems, and they are well suited for applications that are communicationscentric such as Web-based network computing. Another attractive area of mobile agents is processing data over unreliable networks (such as wireless networks). In such an environment, the low reliability network can be used to transfer agents rather than a chunk. of data. The agent can travel to the nodes of the network, collect or process information without the risk of network disconnection, then return home. The publications of this doctorate by published works report on research undertaken in the area of distributed systems with emphasis on network computing paradigms, Web-based distributed computing, and the applications of mobile agents in Web-based distributed computing and wireless computing. The contributions of this collection of related papers can be summarized in four points. First, I have shown how to extend the Web to include computing resources; to illustrate the feasibility of my approach I have constructed a proof of concept implementation. Second, a mobile agent-based approach to Web-based distributed computing, that harness the power of the Web as a computing resource, has been proposed and a system has been prototyped. This, however, means that users will be able to use remote machines to execute their code, but this introduces a security risk. I need to make sure that malicious users cannot harm the remote system. For this, a security policy design pattern for mobile Java code has been developed. Third, a mediator-based approach to wireless client/server computing has been proposed and guidelines for implementing it have been published. This approach allows access to Internet services and distributed object systems from resource-constraint handheld wireless devices such as cellular telephones. Fourth and finally, a mobile agent-based approach to the Wireless Internet has been designed and implemented. In this approach, remote mobile agents can be accessed and used from wireless handheld devices. Handheld wireless devices will benefit greatly from this approach since it overcomes wireless network limitations such as low bandwidth and disconnection, and enhances the functionality of services by being able to operate without constant user input.

Establishing metrics of diversification can calibrate the observed scope of diversity within a lineage and the potential for further phenotypic diversification. There are two potential ways to calibrate differences between phenotypes. The first metric is based on the structure of the network of direct and indirect connections between elements, such as the genes, proteins, enzymes and metabolites that underlie a phenotype. The second metric characterizes the dynamic properties that determine the strength of the interactions among elements, and influence which elements are the most likely to interact. Determining how the connectivity and strength of interactions between elements lead to specific phenotypic variations provides insight into the tempo and mode of observed evolutionary changes. In this dissertation, I proposed and tested hypotheses for how the structure and metabolic flux of a biochemical network delineate patterns of phenotypic variation. I first examined the role of structural properties in shaping observed patterns of carotenoid diversification in avian plumage. I found that the diversification of species-specific carotenoid networks was predictable from the connectivity of the underlying metabolic network. The compounds with the most enzymatic reactions, that were part of the greatest number of distinct pathways, were more conserved across species’ networks than compounds associated with the fewest enzymatic reactions. These results established that compounds with the greatest connectivity act as hotspots for the diversification of pathways between species. Next, I investigated how dynamic properties of biochemical networks influence patterns of phenotypic variation in the concentration and occurrence of compounds. Specifically, I examined if the rate of compound production, known as metabolic flux, is coordinated among compounds in relation to their structural properties. I developed predictions for how different distributions of flux could cause distinct diversification patterns in the concentrations and presence of compounds in a biochemical network. I then tested the effect of metabolic network structure on the concentrations of carotenoids in the plumage of male house finches (Haemorhous mexicanus) from the same population. I assessed whether the structure of a network corresponds to a specific distribution of flux among compounds, or if flux is independent of network structure. I found that flux coevolves with network structure; concentrations of metabolically derived compounds depended on the number of reactions per compound. There were strong correlations between compound concentrations within a network structure, and the strengths of these correlations varied among structures. These findings suggest that changes in network structure, and not independent changes in flux, influence local adaptations in the concentrations of compounds. Lastly, the influence of carotenoid network structure in the evolutionary diversification of compounds across species of birds depends on how the structure of the network itself evolves. To test whether the carotenoid metabolic network structure evolves in birds, I examined the patterns of carotenoid co-occurrence across ancestral and extant species. I found that the same groups of compounds are always gained or lost together even as lineages diverge further from each other. These findings establish that the diversification of carotenoids in birds is constrained by the structure of an ancestral network, and does not evolve independently within a lineage. Taken together, the results of this dissertation establish that local adaptations and the evolutionary diversification of carotenoid metabolism are qualitatively predictable from the structure of an ancestral enzymatic network, and this suggests there is significant structural determinism in phenotypic evolution.

Mathematical modeling of biological phenomena has experienced increasing interest since new high-throughput technologies give access to growing amounts of molecular data. These modeling approaches are especially able to test hypotheses which are not yet experimentally accessible or guide an experimental setup. One particular attempt investigates the evolutionary dynamics responsible for today's composition of organisms. Computer simulations either propose an evolutionary mechanism and thus reproduce a recent finding or rebuild an evolutionary process in order to learn about its mechanism. The quest for evolutionary fingerprints in metabolic and gene-coexpression networks is the central topic of this cumulative thesis based on four published articles. An understanding of the actual origin of life will probably remain an insoluble problem. However, one can argue that after a first simple metabolism has evolved, the further evolution of metabolism occurred in parallel with the evolution of the sequences of the catalyzing enzymes. Indications of such a coevolution can be found when correlating the change in sequence between two enzymes with their distance on the metabolic network which is obtained from the KEGG database. We observe that there exists a small but significant correlation primarily on nearest neighbors. This indicates that enzymes catalyzing subsequent reactions tend to be descended from the same precursor. Since this correlation is relatively small one can at least assume that, if new enzymes are no "genetic children" of the previous enzymes, they certainly be descended from any of the already existing ones. Following this hypothesis, we introduce a model of enzyme-pathway coevolution. By iteratively adding enzymes, this model explores the metabolic network in a manner similar to diffusion. With implementation of an Gillespie-like algorithm we are able to introduce a tunable parameter that controls the weight of sequence similarity when choosing a new enzyme. Furthermore, this method also defines a time difference between successive evolutionary innovations in terms of a new enzyme. Overall, these simulations generate putative time-courses of the evolutionary walk on the metabolic network. By a time-series analysis, we find that the acquisition of new enzymes appears in bursts which are pronounced when the influence of the sequence similarity is higher. This behavior strongly resembles punctuated equilibrium which denotes the observation that new species tend to appear in bursts as well rather than in a gradual manner. Thus, our model helps to establish a better understanding of punctuated equilibrium giving a potential description at molecular level. From the time-courses we also extract a tentative order of new enzymes, metabolites, and even organisms. The consistence of this order with previous findings provides evidence for the validity of our approach. While the sequence of a gene is actually subject to mutations, its expression profile might also indirectly change through the evolutionary events in the cellular interplay. Gene coexpression data is simply accessible by microarray experiments and commonly illustrated using coexpression networks where genes are nodes and get linked once they show a significant coexpression. Since the large number of genes makes an illustration of the entire coexpression network difficult, clustering helps to show the network on a metalevel. Various clustering techniques already exist. However, we introduce a novel one which maintains control of the cluster sizes and thus assures proper visual inspection. An application of the method on Arabidopsis thaliana reveals that genes causing a severe phenotype often show a functional uniqueness in their network vicinity. This leads to 20 genes of so far unknown phenotype which are however suggested to be essential for plant growth. Of these, six indeed provoke such a severe phenotype, shown by mutant analysis. By an inspection of the degree distribution of the A.thaliana coexpression network, we identified two characteristics. The distribution deviates from the frequently observed power-law by a sharp truncation which follows after an over-representation of highly connected nodes. For a better understanding, we developed an evolutionary model which mimics the growth of a coexpression network by gene duplication which underlies a strong selection criterion, and slight mutational changes in the expression profile. Despite the simplicity of our assumption, we can reproduce the observed properties in A.thaliana as well as in E.coli and S.cerevisiae. The over-representation of high-degree nodes could be identified with mutually well connected genes of similar functional families: zinc fingers (PF00096), flagella, and ribosomes respectively. In conclusion, these four manuscripts demonstrate the usefulness of mathematical models and statistical tools as a source of new biological insight. While the clustering approach of gene coexpression data leads to the phenotypic characterization of so far unknown genes and thus supports genome annotation, our model approaches offer explanations for observed properties of the coexpression network and furthermore substantiate punctuated equilibrium as an evolutionary process by a deeper understanding of an underlying molecular mechanism.
Die biologische Zelle ist ein sehr kompliziertes Gebilde. Bei ihrer Betrachtung gilt es, das Zusammenspiel von Tausenden bis Millionen von Genen, Regulatoren, Proteinen oder Molekülen zu beschreiben und zu verstehen. Durch enorme Verbesserungen experimenteller Messgeräte gelingt es mittlerweile allerdings in geringer Zeit enorme Datenmengen zu messen, seien dies z.B. die Entschlüsselung eines Genoms oder die Konzentrationen der Moleküle in einer Zelle. Die Systembiologie nimmt sich dem Problem an, aus diesem Datenmeer ein quantitatives Verständnis für die Gesamtheit der Wechselwirkungen in der Zelle zu entwickeln. Dabei stellt die mathematische Modellierung und computergestützte Analyse ein eminent wichtiges Werkzeug dar, lassen sich doch am Computer in kurzer Zeit eine Vielzahl von Fällen testen und daraus Hypothesen generieren, die experimentell verifiziert werden können. Diese Doktorarbeit beschäftigt sich damit, wie durch mathematische Modellierung Rückschlüsse auf die Evolution und deren Mechanismen geschlossen werden können. Dabei besteht die Arbeit aus zwei Teilen. Zum Einen wurde ein Modell entwickelt, dass die Evolution des Stoffwechsels nachbaut. Der zweite Teil beschäftigt sich mit der Analyse von Genexpressionsdaten, d.h. der Stärke mit der ein bestimmtes Gen in ein Protein umgewandelt, "exprimiert", wird. Der Stoffwechsel bezeichnet die Gesamtheit der chemischen Vorgänge in einem Organismus; zum Einen werden Nahrungsstoffe für den Organismus verwertbar zerlegt, zum Anderen aber auch neue Stoffe aufgebaut. Da für nahezu jede chemische Reaktion ein katalysierendes Enzym benötigt wird, ist davon auszugehen, dass sich der Stoffwechsel parallel zu den Enzymen entwickelt hat. Auf dieser Annahme basiert das entwickelte Modell zur Enzyme-Stoffwechsel-Koevolution. Von einer Anfangsmenge von Enzymen und Molekülen ausgehend, die etwa in einer primitiven Atmosphäre vorgekommen sind, werden sukzessive Enzyme und die nun katalysierbaren Reaktionen hinzugefügt, wodurch die Stoffwechselkapazität anwächst. Die Auswahl eines neuen Enzyms geschieht dabei in Abhängigkeit von der Ähnlichkeit mit bereits vorhandenen und ist so an den evolutionären Vorgang der Mutation angelehnt: je ähnlicher ein neues Enzym zu den vorhandenen ist, desto schneller kann es hinzugefügt werden. Dieser Vorgang wird wiederholt, bis der Stoffwechsel die heutige Form angenommen hat. Interessant ist vor allem der zeitliche Verlauf dieser Evolution, der mittels einer Zeitreihenanalyse untersucht wird. Dabei zeigt sich, dass neue Enzyme gebündelt in Gruppen kurzer Zeitfolge auftreten, gefolgt von Intervallen relativer Stille. Dasselbe Phänomen kennt man von der Evolution neuer Arten, die ebenfalls gebündelt auftreten, und wird Punktualismus genannt. Diese Arbeit liefert somit ein besseres Verständnis dieses Phänomens durch eine Beschreibung auf molekularer Ebene. Im zweiten Projekt werden Genexpressionsdaten von Pflanzen analysiert. Einerseits geschieht dies mit einem eigens entwickelten Cluster-Algorithmus. Hier läßt sich beobachten, dass Gene mit einer ähnlichen Funktion oft auch ein ähnliches Expressionsmuster aufweisen. Das Clustering liefert einige Genkandidaten, deren Funktion bisher unbekannt war, von denen aber nun vermutet werden konnte, dass sie enorm wichtig für das Wachstum der Pflanze sind. Durch Experimente von Pflanzen mit und ohne diese Gene zeigte sich, dass sechs neuen Genen dieses essentielle Erscheinungsbild zugeordnet werden kann. Weiterhin wurden Netzwerke der Genexpressionsdaten einer Pflanze, eines Pilzes und eines Bakteriums untersucht. In diesen Netzwerken werden zwei Gene verbunden, falls sie ein sehr ähnliches Expressionsprofil aufweisen. Nun zeigten diese Netzwerke sehr ähnliche und charakteristische Eigenschaften auf. Im Rahmen dieser Arbeit wurde daher ein weiteres evolutionäres Modell entwickelt, das die Expressionsprofile anhand von Duplikation, Mutation und Selektion beschreibt. Obwohl das Modell auf sehr simplen Eigenschaften beruht, spiegelt es die beobachteten Eigenschaften sehr gut wider, und es läßt sich der Schluss ziehen, dass diese als Resultat der Evolution betrachtet werden können. Die Ergebnisse dieser Arbeiten sind als Doktorarbeit in kumulativer Form bestehend aus vier veröffentlichten Artikeln vereinigt.

The electric grid is undergoing significant changes to meet challenges of improved load control and increased generation from renewables, as well as provision of new services. The main goal of this work is to study the impact of grid’s smartening on the electricity value chain. For this, we built a model to assess investments to come on the grid during the period 2010-2030, both on traditional equipments such as lines and substations and on smartening elements. According to the French example, yearly investments would double on average in the twenty years to come compared to 2010. In the three countries considered in this study, namely France, Italy and Sweden, most investments (between 61% and 76%) occur on the distribution area. Moreover, investments on traditional equipment stand for the lion’s share (68% to 80%) of the total, even if they are partly made possible by the smaller investments on smartening elements, which enable the network to be better controlled. The share of investments on smartening elements is 2.6 to 3.1 times higher on the distribution side than on the transmission side: this denotes the fact that the needed increase in control on the grid is larger on distribution than on transmission. Differences may exist between countries regarding forecasted investments and are mainly due to the number of customers, grid’s size and the chosen generation mix. The study ends with a discussion on the repartition of the value brought by forecasted investments between traditional stakeholders and players that may appear on markets driven by new business models.

This research provides a new approach to analysing the evolutionary nature of social networks that are formed around computer-mediated-communication (CMC) in e-Learning courses. Aspects that have been studied include Online Communities and student communication e-Learning environments. The literature review performed identified weaknesses in the current methods of analyzing CMC activity. A proposed unfied analysis framework (FESNeL) was developed which enables us to explore students' interactions and to test a number of hypotheses. The creation of the framework is discussed in detail along with its major components (e.g. Social Network Analysis and Human Computer Interaction techniques). Furthermore this framework was tested on a case study of an online Language Learning Course. The novelty of this study lies in the investigation of the evolution of online social networks, filling a gap in current research which focuses on specific time stamps (usually the end of the course) when analysing CMC. In addition, the framework uses both qualitative and quantitative methods allowing for a complete assessment of such social networks. Results indicate that FESNeL is a useful methodological framework that can be used to assess student communication and interaction in web-based courses. In addition, through the use of this framework, several characteristic hypotheses were tested which provided useful insights about the nature of learning and communicating online.

Many studies have been conducted regarding network theory and how it can be applied to public transport network. This has led to knowledge on how network indicators relate to the performance of a network and also to insights of how networks can best be extended. Little is known however on how rail bound public transport networks and their network indicators have evolved over time. This would be interesting to know since many metro and other rail bound public transport networks have evolved over a long period of time with extensions being made at different times by different policy makers and stakeholders. This means that there has not been a unified planning process for many of the networks. It would hence be beneficial to get a better picture of how the networks have evolved, when extending the networks or when creating new ones. By creating networks for every year in the development of a rail bound public transport network and then calculate the different network indicators, the evolutionary trends could be found. The networks were created in L-space which means that stations were represented as nodes and the rail connection between stations as edges. To every link in the networks, travel time was attached as weights. This was done in order to make the network indicators more realistic. By assigning geographical coordinates to nodes, indicators such as directness and closeness centrality with respect to geographical distance could be derived. A case study was conducted by applying the methodology to the Stockholm rail bound public transport network. The study period was chosen to be from 1950 up until 2025. 1950 was the year when the Stockholm Metro opened, and the extensions to the network that are decided upon are planned to be completed in 2025. By including the future extensions it was hoped that it could be seen if the future trends are following the trends from the 20th century. Trends regarding the evolution of the network in Stockholm were found. In general it can be said that indicators were relatively high in the first 15-20 years of the study. This was due to the inner city tram network that existed in these years. The tram network was relatively intra-connected with a relatively high average degree, clustering coefficient and connectivity. When the tram network closed down the indicators drastically decreased, after 1971 many of the indicators started to slowly increase due to the additions of new lines and also extensions of already existing ones. Between the year 2000 and 2025, many of the indicators increased substantially, this was partly due to Tvärbanan that connected many older lines creating nodes with a high degree. The fact that the future extensions will lead to an increase in many network indicators (and a decrease in average connectivity) was seen as an indication that the future extensions will accentuate trends that have taken place since the early 1970’s. It was also seen that many of the extensions included in this study will help to develop the network in a way that is in line with the overarching planning principles set by the Stockholm council. The structure of the network consisted of a dense core with branches reaching out to the suburbs in the 1950’s and early 1960’s. In the late 1960’s the network got a radial shape with branches going to the suburbs, no denser core existed in these years. This structure remained relatively unchanged up until the year 2000. After 2000 and up until 2025 a structure emerged in the network with a dense core and also a ring line going around half of the city. This type of structure had been seen in many other rail bound networks around the world.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2006.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Includes bibliographical references (p. 107-114).
The study of protein interactions from the networks point of view has yielded new insights into systems biology [Bar03, MA03, RSM+02, WS98]. In particular, "network motifs" become apparent as a useful and systematic tool for describing and exploring networks [BP06, MKFV06, MSOI+02, SOMMA02, SV06]. Finding motifs has involved either exact counting (e.g. [MSOI+02]) or subgraph sampling (e.g. [BP06, KIMA04a, MZW05]). In this thesis we develop an algorithm to count all instances of a particular subgraph, which can be used to query whether a given subgraph is a significant motif. This method can be used to perform exact counting of network motifs faster and with less memory than previous methods, and can also be combined with subgraph sampling to find larger motifs than ever before -- we have found motifs with up to 15 nodes and explored subgraphs up to 20 nodes. Unlike previous methods, this method can also be used to explore motif clustering and can be combined with network alignment techniques [FNS+06, KSK+03]. We also present new methods of estimating parameters for models of biological network growth, and present a new model based on these parameters and underlying binding domains. Finally, we propose an experiment to explore the effect of the whole genome duplication [KBL04] on the protein-protein interaction network of S. cerevisiae, allowing us to distinguish between cases of subfunctionalization and neofunctionalization.
by Joshua A. Grochow.
M.Eng.

A particular concern within healthcare is the issue of research-informed practice. Failure to translate knowledge efficiently from research into practice potentially has consequences in terms of the quality of care or wasted resources, leading to an inefficient and unproductive health system. Effective techniques and approaches to address this knowledge gap (often called the 'second translational gap') are required. Literature suggests there is no 'magic bullet' to move healthcare research into improved clinical practice. This difficulty is linked, at least in part, to the organisational complexity of health systems including the National Health Service; there are multiple stakeholders, networks and professional and organisational silos. This study draws on data collection and analysis of a healthcare intervention borne from policy aimed specifically at addressing the second translational gap, i.e. moving research into clinical practice effectively and efficiently. The intervention was entitled the 'Collaboration and Leadership in Applied Health Research and Care' (CLAHRC), of which nine examples have been deployed in local health systems. North West London CLAHRC is an appropriate case study as its approach is consciously designed to create collaboration by establishing new networks that span different local health organisations and professions. The study is longitudinal and therefore enables a dynamic perspective that explores the impact of this carefully managed programme of activities on knowledge network evolution within this local context. Using a range of mixed methods, including semi-structured interviews, observation and Social Network Analysis I aimed to uncover how knowledge networks are instigated, how they are successfully developed and also how they are sustained over time, to deliver evidence-based medicine. The findings demonstrate and discuss the process through which a knowledge boundary network evolves and ultimately attains sustainability. It highlights how a mandated, structured inception and continued facilitation leads to increased interaction, a reduction in hierarchy and collaboration across boundaries. The findings are analysed with reference to extant literature and ultimately they contribute to the body of knowledge with regard to boundary network and community development. Finally, this study outlines the implications to future research and in particular the importance of the study to both healthcare practice and policy.

Thesis (MSc)--Stellenbosch University, 2011.
ENGLISH ABSTRACT: In any ecological system, organisms need to interact with each other for their survival. Such interactions form ecological networks which are usually very complex. Nevertheless, they exhibit well de ned patterns; these regularities are often interpreted as products of meaningful ecological processes. As the networks are evolving through time, biological evolution is one of the factors that affects ecological network architecture. In this work, we develop a mathematical model that represents the evolution through time of such ecological interaction networks. The problem is approached by modelling network evolution as a continuous time Markov process, in such a way that the interactions in which a parent species is involved are potentially inherited by its descendant species. This approach allows us to infer ecological parameters and ecological network histories from real-world network data, as well as to simulate ecological networks under our model. While ecologists have long been aware of the in uence of evolutionary processes in shaping ecological networks, we are now able to evaluate the importance of such in uence.
AFRIKAANSE OPSOMMING: In enige ekologiese stelsel benodig organismes wisselwerkings met mekaar ten einde te oorleef. Sulke interaksies vorm ekologiese netwerke wat gewoonlik baie kompleks is maar nogtans goed-gede nieerde patrone vertoon. Hierdie patrone word dikwels geïnterpreteer as die produk van betekenisvolle ekologiese prosesse. Aangesien die netwerke met die verloop van tyd ontwikkel, is biologiese ewolusie een van die faktore wat ekologiese netwerkargitektuur beïnvloed. In hierdie studie ontwikkel ons 'n wiskundige model wat die ewolusie van sulke ekologiese interaksienetwerke voorstel. Die probleem word benader deur netwerkewolusie as 'n kontinue-tyd Markov-proses te modelleer, op so 'n manier dat die interaksies waarin 'n voorouerspesie betrokke is potensieel oorerf kan word deur die afstammelingspesies. Hierdie benadering laat ons toe om ekologiese parameters en ekologiese netwerkgeskiedenisse vanuit regte-wêreld data af te lei, sowel as om ekologiese netwerke onder ons model te simuleer. Alhoewel ekoloë al lank reeds bewus is van die invloed wat ewolusionêre prosesse het op die vorming van ekologiese netwerke, is ons nou in staat om die belangrikheid van hierdie invloed te evalueer.

Background Recurrence and predictability of evolution are thought to reflect the correspondence between genomic and phenotypic dimensions of organisms, and the connectivity in deterministic networks within these dimensions. Direct examination of the correspondence between opportunities for diversification imbedded in such networks and realized diversity is illuminating, but is empirically challenging because both the deterministic networks and phenotypic diversity are modified in the course of evolution. Here we overcome this problem by directly comparing the structure of a “global” carotenoid network – comprising of all known enzymatic reactions among naturally occurring carotenoids – with the patterns of evolutionary diversification in carotenoid-producing metabolic networks utilized by birds. Results We found that phenotypic diversification in carotenoid networks across 250 species was closely associated with enzymatic connectivity of the underlying biochemical network – compounds with greater connectivity occurred the most frequently across species and were the hotspots of metabolic pathway diversification. In contrast, we found no evidence for diversification along the metabolic pathways, corroborating findings that the utilization of the global carotenoid network was not strongly influenced by history in avian evolution. Conclusions The finding that the diversification in species-specific carotenoid networks is qualitatively predictable from the connectivity of the underlying enzymatic network points to significant structural determinism in phenotypic evolution.

Les modifications post-traductionnelles (PTM) sont des modifications chimiques des protéines qui permettent à la cellule de réguler finement ses fonctions ainsi que de coder et d’intégrer des signaux environnementaux. Les progrès récents en ce qui a trait aux techniques expérimentales et bioinformatiques nous ont permis de determiner les profils de PTM pour des protéomes entiers ainsi que d’identifier les molécules qui sont responsables d’ « écrire » ou d’« effacer » ces PTM. Avec ces donnés, il a été possible de commencer à definir des réseaux de régulation cellulaire par PTM. Ici, nous avons étudié l’évolution de ces réseaux pour mieux comprendre comment ils peuvent contribuer à expliquer la complexité et la diversité des organismes ainsi que pour mieux comprendre leurs mecanismes d’action. Avant tout, nous avons abordé la question de comment les réseaux de régulation des PTM peuvent être recablés après un évenement de duplication des gènes en étudiant comment le réseau de phosphorégulation de la levure bourgeonnante a été récablé après un évenement de duplication complète du génome qui a eu lieu il y a 100 milions d’années. Nos résultats mettent en évidence le rôle de la duplication des gènes comme mécanisme clé pour l’innovation et la complexification des réseaux de régulation par PTM. Par la suite, nous avons abordé la question de comment les PTM peuvent contribuer à la diversité des organismes en comparant les profils de phosphorylation de l’homme et de la souris. Nous avons trouvé des différences substantielles dans les profils de PTM de ces deux espèces qui ont le potentiel d’expliquer, au moins en partie, les différences phénotypiques observées entre eux. Nous avons aussi trouvé des évidences qui supportent l’idée que les PTM peuvent « sauter » vers des nouvelles localisations et quand même réguler les mêmes fonctions biologiques. Ce phénomène doit être pris en considération dans les comparaisons des profils de PTM qui appartiennent à des espèces différentes, pour éviter de surestimer la divergence causée par la régulation par les PTM. Enfin, nous avons investigué comment plusieures PTM alternatives pour un même residu pouvent interagir pour réguler des fonctions cellulaires. Nous avons examiné deux des PTM les plus connus, la phosphorylation et la O-GlcNAcylation, qui modifient les sérines et les thréonines, et nous avons étudié les mécanismes potentiels d’interaction entre ces deux PTM. Nos résultats supportent l’hypothèse que ces deux PTM contrôlent plusieurs fonctions biologiques plutôt qu’une seule fonction. Globalement, les résultats présentés dans cette thèse permettent d’élucider les dynamiques évolutives, les mécanismes de fonctionnement et les implications biologiques des PTM.
Post-translational modifications (PTMs) are chemical modification of proteins that allow the cell to finely tune its functions as well as to encode and integrate environmental signals. The recent advancements in the experimental and bioinformatic techniques have allowed us to determine the PTM profiles of entire proteomes as well as to identify the molecules that write or erase PTMs to/from each protein. This data have made possible to define cellular PTM regulatory networks. Here, we study the evolution of these networks to get new insights about how they may contribute to increase organismal complexity and diversity and to better understand their molecular mechanisms of functioning. We first address the question of how and to which extent a PTM network can be rewired after a gene duplication event, by studying how the budding yeast phosphoregulatory network was rewired after a whole genome duplication event that occurred 100 million years ago. Our results highlight the role of gene duplication as a key mechanism to innovate and complexify PTM regulatory networks. Then, we address the question of how PTM networks may contribute to organismal diversity by comparing the human and mouse phosphorylation profiles. We find that there are substantial differences in the PTM profiles of these two species that have the potential to explain, at least in part, the phenotypic differences observed between them. Moreover, we find evidence supporting the idea that PTMs can jump to new positions during evolution and still regulate the same biological functions. This phenomenon should be taken into account when comparing the PTM profiles of different species, in order to avoid overestimating the divergence in PTM regulation. Finally, we investigate how multiple and alternative PTMs that affect the same residues interact with each other to control proteins functions. We focus on two of the most studied PTMs, protein phosphorylation and O-GlcNAcylation, that affect serine and threonine residues and we study their potential mechanisms of interactions in human and mouse. Our results support the hypothesis that these two PTMs control multiple biological functions rather than a single one. Globally this work provides new findings that elucidate the evolutionary dynamics, the functional mechanisms and the biological implications of PTMs.