The evolution of plant genomes is notable for manifesting a cycle of whole genome doubling, fractionation (gradual loss of redundant genes) and speciation. The thesis is based on a branching process model of the doubling and fractionation process, integrated with a standard model of sequence divergence. The immediate application of this work is to account for the distribution of sequence similarity for duplicate gene pairs, both within plant genomes and between two related plant genomes in terms of a cycle of polyploidization, fractionation and speciation. We derive a mixture distribution for duplicate gene pair similarities generated by speciation and/or repeated episodes of polyploidization. We account not only for the timing of these events in terms of local modes or peaks of the component distributions, but also their volume, or amplitude, and variance. We outline how to infer the parameters of the model. We illustrate with analyses of the distribution of homolog similarities in a number of plant families: Brassicaceae, Solanaceae and Malvaceae. To our knowledge, this is the first method to account for the volume of the component normals of a distribution of similarities, preliminary to an evolutionarily meaningful inference procedure. In addition, we solve the problem of identifying the ploidy level of a series of two or three polyploidizations by invoking the observed and predicted gene triple profiles for each model, i.e., by calculating the probability of the four types of triple with origins in one or the other event, or both.

Major efforts to sequence cancer genomes are now occurring throughout the world. Though the emerging data from these studies are illuminating, their reconciliation with epidemiologic and clinical observations poses a major challenge. Here we present mathematical models that begin to address this challenge. First we present a model of accumulation of driver and passenger mutations during tumor progression and derive a formula for the number of driver mutations as a function of the total number of mutations in a tumor. Fitting this formula to recent experimental data, we were able to calculate the selective advantage provided by a typical driver mutation. Second, we performed a quantitative analysis of pancreatic cancer metastasis genetic data. The results of this analysis define a broad time window for detection of pancreatic cancer before metastatic dissemination. Finally, we model the evolution of resistance to targeted cancer therapy. We apply our model to experimental data on the response to panitumumab, targeted therapy against colorectal cancer. Our modeling suggested that cells resistant to therapy were likely present in patients’ tumors prior to the start of therapy. / Mathematics

Algumas metapopulações de espécies, como formigas, vivem em colônias que crescem durante algum tempo e depois colapsam. Após o colapso poucos indivíduos sobrevivem. Esses indivíduos se dispersam tentando fazer novas colônias que podem ou não se estabelecer dependendo do ambiente que encontram. Recentemente, Schinazi (2015) usou cadeias de nascimento e morte em ambientes aleatórios para modelar tais populações, e mostrou que a dispersão aleatória é uma boa estratégia para a sobrevivência da população. Nesta tese, introduzimos outros modelos estocásticos de colonização e colapso para os quais consideramos restrições espaciais e diferentes tipos de colapsos. Obtemos para esses novos modelos condições de sobrevivência e extinção. Debatemos algumas situações nas quais a dispersão nem sempre é uma boa estratégia de sobrevivência. Além disso, discutimos a relação destes modelos com outros conhecidos na literatura. Técnicas de percolação, acoplamento e comparação com processos de ramificação convenientemente definidos são usadas para obter os resultados aqui estabelecidos. / Some metapopulations, such as ants, live in colonies that grow for a while and then collapse. Upon collapse, very few individuals survive. These individuals disperse, trying to establish new colonies that may or may not settle, depending on the environment they encounter. Recently, Schinazi (2015) used birth and death chains in random environments to model such populations, and showed that random dispersion is a good strategy for the survival of the population. In this thesis, we introduce other stochastic models of colonization and collapse for which we consider spatial constraints and different kinds of collapse. We obtain conditions for survival and extinction in these new models. We discuss some situations in which dispersion is not always a good survival strategy. In addition, we discuss the relation of these models to others known in the literature. Percolation and coupling techniques and comparison with suitably defined branching processes are used to obtain the results set forth herein.

Algumas metapopulações de espécies, como formigas, vivem em colônias que crescem durante algum tempo e depois colapsam. Após o colapso poucos indivíduos sobrevivem. Esses indivíduos se dispersam tentando fazer novas colônias que podem ou não se estabelecer dependendo do ambiente que encontram. Recentemente, Schinazi (2015) usou cadeias de nascimento e morte em ambientes aleatórios para modelar tais populações, e mostrou que a dispersão aleatória é uma boa estratégia para a sobrevivência da população. Nesta tese, introduzimos outros modelos estocásticos de colonização e colapso para os quais consideramos restrições espaciais e diferentes tipos de colapsos. Obtemos para esses novos modelos condições de sobrevivência e extinção. Debatemos algumas situações nas quais a dispersão nem sempre é uma boa estratégia de sobrevivência. Além disso, discutimos a relação destes modelos com outros conhecidos na literatura. Técnicas de percolação, acoplamento e comparação com processos de ramificação convenientemente definidos são usadas para obter os resultados aqui estabelecidos. / Some metapopulations, such as ants, live in colonies that grow for a while and then collapse. Upon collapse, very few individuals survive. These individuals disperse, trying to establish new colonies that may or may not settle, depending on the environment they encounter. Recently, Schinazi (2015) used birth and death chains in random environments to model such populations, and showed that random dispersion is a good strategy for the survival of the population. In this thesis, we introduce other stochastic models of colonization and collapse for which we consider spatial constraints and different kinds of collapse. We obtain conditions for survival and extinction in these new models. We discuss some situations in which dispersion is not always a good survival strategy. In addition, we discuss the relation of these models to others known in the literature. Percolation and coupling techniques and comparison with suitably defined branching processes are used to obtain the results set forth herein.

<p> Techniques from algebra and matrix theory are employed to study the total progeny of a multitype branching process from the point of probability generating functions. A result for the total progeny of different types of individuals having identical offspring distribution is developed, which extends the classic Dwass formula from single case to multitype case. An example with Poisson distributed offspring having different distributions of children is given to illustrate that total progeny does not preserve similar structure as Dwass' formula in general.</p> / Thesis / Master of Science (MSc)

Estudamos um sistema de partículas a tempo discreto cuja dinâmica é a seguinte. Considere que no instante inicial sobre cada inteiro não negativo há uma partícula, inicialmente inativa. A partícula da origem é ativada e instantaneamente ativa um conjunto aleatório contíguo de partículas que estão a sua direita. Como regra, no instante seguinte ao que foi ativada, cada partícula ativa realiza esta mesma dinâmica de modo independente de todo o resto. Dizemos que o processo sobrevive se em qualquer momento sempre há ao menos uma partícula ativa. Chamamos este processo de Firework, associando a dinâmica de ativação de uma partícula inativa a uma infecção ou explosão. Nosso interesse é estabelecer se o processo tem probabilidade positiva de sobrevivência e apresentar limites para esta probabilidade. Isto deve ser feito em função da distribuição da variável aleatória que define o raio de ação de uma partícula. Associando o processo de ativação a uma infecção, podemos pensar este modelo como um modelo epidêmico. Consideramos também algumas variações dessa dinâmica. Dentre elas, variantes com partículas distribuídas sobre a semirreta dos reais positivos (nesta vertente, existem condições para as distâncias entre partículas consecutivas) e também com as partículas distribuídas sobre vértices de árvores. Estudamos também para esses casos a transição de fase e probabilidade de sobrevivência. Nesta variante os resultados obtidos são funções da sequência de distribuições dos alcances das explosões e da estrutura dos lugares onde se localizam as partículas. Consideramos também variações do modelo onde cada partícula ao ser ativada, permanece ativa durante um tempo aleatório e nesse período emite explosões que ocorrem em instantes aleatórios. / We studied a discrete time particle system whose dynamic is as follows. Consider that at time zero, on each non-negative integer, there is a particle, initially inactive. A particle which is placed at origin is activated and instantly activates a contiguous random set of particles that is on its right. As a rule, the next moment to what it has been activated, each active particle carries the same behavior independently of the rest. We say that the process survives if the amount of particles activated along the process is infinite. We call this the Firework process, associating the activation dynamic of a particle to an infection or explosion process. Our interest is to establish whether the process has positive probability of survival and to present limits to this probability. This is done according to the distribution random variable that defines the radius of infection of each active particle, Associating the activation process to an infection, we think this model as a model epidemic. We also consider some variations of this dynamic. Among them, variants with particles distributed over the half line (there are conditions for the distances between consecutive particles) and also with particles distributed over the vertices of a tree. We studied phase transitions and the correspondent survival probability. In this variant the results depend on the sequence of probability distributions for the range of the explosions and on the particles displacement. We also consider a variation where each particle after activated, remains active during a random time period emitting explosions that occur in random moments.

Estudamos um sistema de partículas a tempo discreto cuja dinâmica é a seguinte. Considere que no instante inicial sobre cada inteiro não negativo há uma partícula, inicialmente inativa. A partícula da origem é ativada e instantaneamente ativa um conjunto aleatório contíguo de partículas que estão a sua direita. Como regra, no instante seguinte ao que foi ativada, cada partícula ativa realiza esta mesma dinâmica de modo independente de todo o resto. Dizemos que o processo sobrevive se em qualquer momento sempre há ao menos uma partícula ativa. Chamamos este processo de Firework, associando a dinâmica de ativação de uma partícula inativa a uma infecção ou explosão. Nosso interesse é estabelecer se o processo tem probabilidade positiva de sobrevivência e apresentar limites para esta probabilidade. Isto deve ser feito em função da distribuição da variável aleatória que define o raio de ação de uma partícula. Associando o processo de ativação a uma infecção, podemos pensar este modelo como um modelo epidêmico. Consideramos também algumas variações dessa dinâmica. Dentre elas, variantes com partículas distribuídas sobre a semirreta dos reais positivos (nesta vertente, existem condições para as distâncias entre partículas consecutivas) e também com as partículas distribuídas sobre vértices de árvores. Estudamos também para esses casos a transição de fase e probabilidade de sobrevivência. Nesta variante os resultados obtidos são funções da sequência de distribuições dos alcances das explosões e da estrutura dos lugares onde se localizam as partículas. Consideramos também variações do modelo onde cada partícula ao ser ativada, permanece ativa durante um tempo aleatório e nesse período emite explosões que ocorrem em instantes aleatórios. / We studied a discrete time particle system whose dynamic is as follows. Consider that at time zero, on each non-negative integer, there is a particle, initially inactive. A particle which is placed at origin is activated and instantly activates a contiguous random set of particles that is on its right. As a rule, the next moment to what it has been activated, each active particle carries the same behavior independently of the rest. We say that the process survives if the amount of particles activated along the process is infinite. We call this the Firework process, associating the activation dynamic of a particle to an infection or explosion process. Our interest is to establish whether the process has positive probability of survival and to present limits to this probability. This is done according to the distribution random variable that defines the radius of infection of each active particle, Associating the activation process to an infection, we think this model as a model epidemic. We also consider some variations of this dynamic. Among them, variants with particles distributed over the half line (there are conditions for the distances between consecutive particles) and also with particles distributed over the vertices of a tree. We studied phase transitions and the correspondent survival probability. In this variant the results depend on the sequence of probability distributions for the range of the explosions and on the particles displacement. We also consider a variation where each particle after activated, remains active during a random time period emitting explosions that occur in random moments.

In this paper, we focus on the goodness of fit test for self-similar property of two well-known processes: the fractional Brownian motion and the fractional autoregressive integrated moving average process. The Hurst parameter of the self-similar process is estimated by the embedding branching process method proposed by Jones and Shen (2004). The goodness of fit test for self-similarity is based on the Pearson chi-square test statistic. We approximate the null distribution of the test statistic by a scaled chi-square distribution to correct the size bias problem of the conventional chi-square distribution. The scale parameter and degrees of freedom of the test statistic are determined via regression method. Simulations are performed to show the finite sample size and power of the proposed test. Empirical applications are conducted for the high frequency financial data and human heart rate data.

<p>The first part of the thesis focuses on the development of Bayesian modeling motivated by geophysics applications. In Chapter 2, we model the frequency of pyroclastic flows collected from the Soufriere Hills volcano. Multiple change points within the dataset reveal several limitations of existing methods in literature. We propose Bayesian hierarchical models (BBH) by introducing an extra level of hierarchy with hyper parameters, adding a penalty term to constrain close consecutive rates, and using a mixture prior distribution to more accurately match certain circumstances in reality. We end the chapter with a description of the prediction procedure, which is the biggest advantage of the BBH in comparison with other existing methods. In Chapter 3, we develop new statistical techniques to model and relate three complex processes and datasets: the process of extrusion of magma into the lava dome, the growth of the dome as measured by its height, and the rockfalls as an indication of the dome's instability. First, we study the dynamic Negative Binomial branching process and use it to model the rockfalls. Moreover, a generalized regression model is proposed to regress daily rockfall numbers on the extrusion rate and dome height. Furthermore, we solve an inverse problem from the regression model and predict extrusion rate based on rockfalls and dome height.</p><p>The other focus of the thesis is adaptive Markov chain Monte Carlo (MCMC) method. In Chapter 4, we improve upon the Wang-Landau (WL) algorithm. The WL algorithm is an adaptive sampling scheme that modifies the target distribution to enable the chain to visit low-density regions of the state space. However, the approach relies heavily on a partition of the state space that is left to the user to specify. As a result, the implementation and the use of the algorithm are time-consuming and less automatic. We propose an automatic, adaptive partitioning scheme which continually refines the initial partition as needed during sampling. We show that this overcomes the limitations of the input user-specified partition, making the algorithm significantly more automatic and user-friendly while also making the performance dramatically more reliable and robust. In Chapter 5, we consider the convergence and autocorrelation aspects of MCMC. We propose an Exploration/Exploitation (XX) approach to constructing adaptive MCMC algorithms, which combines adaptation schemes of distinct types. The exploration piece uses adaptation strategies aiming at exploring new regions of the target distribution and thus improving the rate of convergence to equilibrium. The exploitation piece involves an adaptation component which decreases autocorrelation for sampling among regions already discovered. We demonstrate that the combined XX algorithm significantly outperforms either original algorithm on difficult multimodal sampling problems.</p> / Dissertation

Pathogen evolution poses a significant challenge to public health, as efforts to control the spread of infectious diseases struggle to keep up with a shifting target. To better understand this adaptive process, we turn to mathematical modelling. Specifically, we use multi-type branching processes to describe a pathogen's stochastic spread among members of a host population or growth within a single host. In each case, there is potential for new pathogen strains with different characteristics to arise through mutation. We first develop a specific model to study the emergence of a newly introduced infectious disease, where the pathogen must adapt to its new host or face extinction in this population. In an extension of previous models, we separate the processes of host-to-host contacts and disease transmission, in order to consider each of their contributions in isolation. We also allow for an arbitrary distribution of host contacts and arbitrary mutational pathways/rates among strains. This framework enables us to assess the impact of these various factors on the chance that the process develops into a large-scale epidemic. We obtain some intriguing results when interpreted in a biological context. Secondly, motivated by a desire to investigate the time course of pathogen evolutionary processes more closely, we derive some novel theoretical results for multi-type branching processes. Specifically, we obtain equations for: (1) the distribution of waiting time for a particular type to arise; and (2) the distribution of population numbers over time, conditioned on a particular type not having yet appeared. A few numerical examples scratch the surface of potential applications for these results, which we hope to develop further. / Thesis (Master, Mathematics & Statistics) -- Queen's University, 2010-07-28 11:43:22.984