background
XI Um dia de Biometria

11/10/2019 - 11/10/2019

Botucatu

Descrição Resumida

Um dia de Biometria é um evento anual, organizado por pós-graduandos e docentes do Programa de Pós-Graduação em Biometria do Instituto de Biociências da UNESP de Botucatu.
O evento tem como objetivo principal abordar pesquisas recentes em Bioestatística e Modelagem Matemática de Biossistemas, buscando promover discussões e interações entre discentes, docentes e profissionais da área.
Em sua 11ª edição, os discentes do Programa de Pós-Graduação em Biometria e demais interessados poderão apresentar seus trabalhos em forma de pôster, de modo a divulgar e promover a pesquisa que vêm desenvolvendo, e assim, obter contribuições significativas.

O púbico alvo do evento XI Um dia de Biometria são:
- Discentes do Programa de Pós-Graduação em Biometria;
- Discentes da graduação e outros programas de pós-graduação;
- Docentes, pesquisadores, e profissionais da área;
- Demais interessados;

 

Programação

The power of complex networks.
Francisco Aparecido Rodrigues, ICMC, USP. 

Complex networks are everywhere, from social networks to the biological interactions in our body. These networks represent the organization of complex systems and are the medium in which information, diseases or failures propagate. In this talk, we will show how Network Science has been used to model different problems in our society, including some biological applications. Our last results related to modelling of epidemic processes and the synchronization of biological oscillators will be presented. Applications related to traffic flow in cities, protein-protein interactions and the diagnosis of mental disorders will also be discussed.  Finally, we will discuss the future of Network Science and the current challenges, including issues in systems biology.

 

Quasi U-statistics and substitution models for tests of homogeneity of groups of DNA sequences.
Hildete Prisco Pinheiro, IMECC, UNICAMP.

 

One of the goals in genetic studies is the comparison of groups of genetic sequences. The main purpose of this work is to test homogeneity among groups of DNA sequences incorporating substitution models on the distance between sequences. These models describe the stochastic process of the mutations through time. Diversity measures between sequences may be based on Hamming distances incorporating evolutionary models of substitution. A test statistic based on quasi U-statistics is proposed, relaxing the assumption of independence between sequences. An application to real data set illustrates the test procedure.

 

 

Generating p-rep designs.
Renata Alcarde Sermarini, ESALQ, USP.

 

An important objective in plant breeding is to maximize the genetic gain arising from genetic selection. This requires efficient designs in circumstance in which there are limitations on the resources, particularly in the early phase of a breeding program. In general, the number of new lines, or test lines, is large and the genetic material available for the experiment is not enough to allow the researcher to replicate all the test lines in the same experiment, and as well the space in the field can be restricted. For these reasons p-rep designs are widely used. The search for an optimal design is one way to identify a design that is as efficient as possible. However, the designs identified depend on assumptions, such as the design model and the criterion search. This study evaluates the differences in genetic selection, as measured by the realized genetic gain, between optimized resolved and non-resolved p-rep designs for early generation plant breeding experiments under (i) different assumptions about treatment effects, fixed or random, (ii) different values for the nongenetic parameters in the design model, and (iii) different optimality criteria, C- and D-optimality. Two simulation studies were conduct in order to assess the effect of these design characteristics on the realized genetic gain and, consequently, on the selection. The first, a large experiment, showed small differences between the design scenarios and search criteria for a given data scenario (less than 4%). The second, a smaller experiment, showed higher differences in the presence of low data genetic variance (up to 13%). Further, a third simulation study was conducted to evaluate, for the smaller experiment, the influence of the p, the percentage of treatment replication on the realized genetic gain. It is clear that increasing the number of replicated treatments, rather than increasing the realized genetic gain, improves the precision with which the realized genetic gain is estimated. Strategies for identifying a design for an experiment will be outlined.

 

 

A family of generalized linear models for repeated measures with normal and conjugate random effects.
Clarice Garcia Borges Demétrio, ESALQ, USP.

 

Non-Gaussian outcomes are often modeled using members of the so-called exponential family. Notorious members are the Bernoulli model for binary data, leading to logistic regression, and the Poisson model for count data, leading to Poisson regression. Two of the main reasons for extending this family are (1) the occurrence of overdispersion, meaning that the variability in the data is not adequately described by the models, which often exhibit a prescribed mean-variance link (Hinde and Demétrio 1998, Computational Statistics and Data Analysis 27, 151-170), and (2) the accommodation of hierarchical structure in the data, stemming from clustering in the data which, in turn, may result from repeatedly measuring the outcome, for various members of the same family, etc. The first issue is dealt with through a variety of overdispersion models, such as, for example, the beta-binomial model for grouped binary data and the negative-binomial model for counts. Clustering is often accommodated through the inclusion of random subject-specific effects. Though not always, one conventionally assumes such random effects to be normally distributed. While both of these phenomena may occur simultaneously, models combining them are uncommon. This paper proposes a broad class of generalized linear models accommodating overdispersion and clustering through two separate sets of random effects (Molenberghs, Verbeke and Demétrio 2007, LIDA, 13, 513-531; Molenberghs, Verbeke and Demétrio 2017, SORT, 41:191:254). We place particular emphasis on so-called conjugate random effects at the level of the mean for the first aspect and normal random effects embedded within the linear predictor for the second aspect, even though our family is more general. The binary, count, and time-to-event cases are given particular emphasis. Apart from model formulation, we present an overview of estimation methods, and then settle for maximum likelihood estimation with analytic-numerical integration. Implications for the derivation of marginal correlations functions are discussed. The methodology is applied to data from a study in epileptic seizures, a clinical trial in toenail infection named onychomycosis, and survival data in children with asthma (Molenberghs et al, 2010, Statistical Science, 25: 325–347).

 

 

Modelos para dinâmica de populações com estrutura de estágios: como mudança de temperatura afeta populações de insetos?
Renato Mendes Coutinho, CCMC, UFABC.

 

Entender como a variação de temperatura influencia populações ecológicas é um objetivo central face ao aquecimento global. Animais que não regulam sua temperatura interna (ectotermos) e cujo desenvolvimento é fortemente afetado por ela são especialmente vulneráveis, e dentre estes os insetos são os principais representantes, tanto por sua abundância e diversidade quanto por seu papel ecológico e importância econômica.
Uma abordagem para modelar esse problema é descrever explicitamente sua história de vida, ou seja, seus estágios larvais e adulto. Para isso, construímos modelos estruturados por idade ou por estágio de desenvolvimento, via equações diferenciais parciais. Uma forma de simplificar tais modelos é derivando, a partir das EDPs básicas, sistemas de equações diferenciais com atraso temporal equivalentes. Finalmente, é necessário ainda elucidar como as taxas relevantes-fecundidade, mortalidade e desenvolvimento-dependem da temperatura, o que pode ser feito a partir de hipóteses gerais sobre a biologia desses seres vivos, e empiricamente em sistemas específicos.
É ainda uma questão em aberto como a temperatura afeta, em geral, as interações ecológicas, embora muitos estudos recentes busquem quantificar esses efeitos, tanto em laboratório quanto no campo.

Regulamento das Inscrições

- As inscrições são gratuitas e devem ser realizadas pelo site do evento.
- Serão permitidas as inscrições de discentes e docentes de Graduação, Pós-Graduação, Pesquisadores e Profissionais interessados.
- As inscrições podem ser realizadas até o dia do evento.

Normas para Submissão dos Trabalhos

Instruções aos autores para elaboração dos resumos 

Os autores são convidados a submeter um resumo dentro das especificações seguintes:
- O resumo poderá ser escrito em português ou inglês.
- O tamanho do resumo não deve exceder uma folha A4, incluindo título, autores, afiliações, texto, e palavras-chave. 
- Margens superior, inferior, esquerda e direita devem ter 2,5 cm.
- Título centralizado, em negrito, letras maiúsculas e fonte Times New Roman 13.
- Nome de cada autor centralizado, em negrito, itálico e fonte Times New Roman 11, seguidos da filiação e e-mail, centralizados, em itálico e fonte Times New Roman 11.
- O Título e Nomes dos autores devem ter espaçamento simples.
- O texto do resumo deve estar justificado e ter de 150 a 400 palavras, em fonte Times New Roman 11 e espaçamento de 1,5 entre linhas.
- O termo palavras-chave em negrito e letras maiúsculas.  Máximo de 5 palavras-chave em itálico e fonte Times New Roman 11. 
- O espaço de uma linha deverá separar o título, autores, afiliações, texto e palavras-chave.

Modelo de resumo, clique aqui: Download Template.

Submissão de resumos

O inicio da submissão dos resumos será no dia 15/07/2019 e o prazo final no dia 16/09/2019.

Prorrogação da data para envio de trabalhos: 24/09/2019.

 

Aceite

A decisão final sobre o aceite do resumo bem como a sessão para apresentação do trabalho será comunicada ao autor correspondente até 01/10/2019. 

Apresentação

Os trabalhos aceitos deverão ser apresentados em forma de pôster.
O pôster deve respeitar a medida de: 90cm de largura por 100 cm de altura.
Além do pôster, os autores deverão preparar uma apresentação de no máximo 2 slides para uma breve apresentação oral de 2 minutos.
Essa apresentação terá o objetivo de caracterizar cada pôster e despertar o interesse dos participantes para conhecerem os trabalhos.
Os pôsteres serão apresentados no evento durante os coffee-breaks.
Os pôsteres que permanecerem nos painéis após o prazo de desmontagem serão removidos pela organização e descartados.

Organização

Apoio

Local do Evento

Anfiteatro do IB - Unesp - Botucatu.