In this talk we will present two unrelated methods for constructing networks from molecular sequence data
and from metabolic profiles.

Phylogenetic networks are a generalization of phylogenetic trees that permit the representation of conflicting signal or alternative phylogenetic histories. Networks can provide a useful tool for phylogenetic analysis when the underlying evolutionary history is non treelike. For example, recombination, hybridization, and lateral gene transfer can all lead to histories that are not adequately represented by a single tree. In the first part

The problem we address is text indexing for approximate matching. We consider that we are given a text T which undergoes some preprocessing to generate an index. We can later query this index to identify the places where a string occurs up to a certain number of allowed substitutions k. We present a structure for indexing which occupies space O(n log^k n) in the average case, independent of alphabet size, n being the text size. This structure support searching in O(m^{k+1}) time, for patterns of size m, again independent of alphabet size.

In this talk we will focus on biclustering, a technique that has recently shown to be remarkably effective in a variety of applications in biological data analysis and other data mining tasks. The importance of biclustering in the identification of groups of genes with coherent expression patterns (in a subset of the experimental conditions), and its advantages (when compared to clustering) in the discovery of local expression patterns has been extensively studied and documented.

Over the past years, several scheduling techniques have been proposed for high level synthesis. However, the main focus of the research in this area has been in the development of optimal algorithms for small problems. Since the scheduling problem as been shown to be NP-complete, simpler algorithms must be investigated for scheduling large problems. In this sense, it will be shown that by using simple algorithms it is possible to acheive solutions close to optimal in a very short time.

A apresentação centra-se em emparelhamento directamente no texto comprimido. Será abordada a técnica de compressão "códigos de redundância mínima" (Códigos de Huffman), dando especial interesse à codificação de "Tagged Huffman Coding". Baseado nesta codificação, serão apresentadas duas formas de pesquisa no texto comprimido.

RNA is an essential element in every living organism where it plays crucial roles. Since RNA structure is strictly related to its function, it is important to find common RNA motifs in order to find functional similarities. We will present the method proposed by Backofen and Siebert to solve the exact common sequence/pattern problem, which was unsolved until 2004.

Logistic regression is one of the workhorses of statistical learning and it has been used and studied for several decades. In this talk I will focuss on two recent advances in logistic regression:

Todos nós, portugueses, seguimos com atenção o concurso de colocação dos docentes dos ensinos pré-escolar, básico e secundário, ano lectivo de 2004/2005. Pelo que na altura veio a público, ficou claro que no âmago da questão encontrava-se um problema de natureza técnica, matemática, que não fora devidamente valorizado pelo Ministério da Educação. Nesta palestra, pretendemos identificar as questões de índole matemática que se levantam num concurso para colocação de docentes.

Após a sequenciação dos genomas de diversos organismos, passou-se à fase de anotação dos genes, tendo muita desta informação ficado disponível para ser processada e transformada em conhecimento. Neste contexto, é especialmente importante o estudo dos mecanismos de interacções entre genes, ou seja, o estudo das redes de regulação genética.

The purpose of this project was to provide email indexing for kmail, an open-source email client. As underlying indexing structure, inverted files were used. The project is broadly divided into a core library and its use by kmail. This talk will focus mostly on the core library, its design and implementation. The kmail integration will be presented as a demonstration.