Modelo com mistura de multinomiais aplicado à identificação de proteínas similares

Abstract

The proteins are important molecules from the cells, whereas they take part since the construction of cell´s framing until the transmission of the genetic information between the generations. A protein can be characterized by its function and its function is determined by the sequence of amino acids that determines its structure. To determined the protein's function is important, for instance, in a research about the cure of diseases or searching for new drugs. In this research we use a bayesian statistical methodology with mixture of multinomial and latent variables to identify proteins with similar function. We use simulations to verify the performance of the statistical model for identifying the similar proteins. At the end we apply the modeling to a real data set.