Cluster: um software para auxílio em estudos de dados biológicos

Abstract

The ever increasing availability of biological data gives rise to two problems: (i) data storage and management and (ii) the extraction of useful information from these data. The latter problem is one of the main challenges in computational biology, and requires the development of tools and methods capable of transforming all these heterogeneous data into biological knowledge. Part of this knowledge involves determining variations in gene expression on biological data. Studies on biological data have contributed to the development of new techniques in agriculture, animal farming, in the treatment of diseases and in the development of policies for the preservation of endangered animal and plant species. Thus, this paper proposes a software, named Cluster, to assist research on genetic diversity. Cluster acts directly on the feature selection step of the classification problem. Cluster is able to optimize the quantity and quality of the features used to group individuals. The simple interface of the Cluster software helps its configuration and the presentation of clear results. The software is tested on databases with different properties. The specificity, sensitivity, efficiency and accuracy of the classification are metrics used to validate the feature selection mechanism proposed in Cluster. Tests performed on the software include: the determination of alleles for distinguishing sea turtles and their hybrids; the determination of genomic features for classification gastric cancer tissue and determination of morphological features for classification wheat seeds.