Ranking genes based on topics

The vast majority of existing methods for ranking gene datasets, do not include or take into account in their exported results other information that might accompany the genes, such as specie or terms. Given a higher order biological data set, we propose a methodology based on multilinear algebra for ranking genes across multiple dimensions. We apply PARAFAC decomposition on a Gene Ontology dataset (GO data) for multiple species and reveal interesting experimental results that provide to the user more information than other consisted methods.