LeaderGene: A Fast Data-mining Tool for Molecular Genomics

Abstract

DNA microarrays are one of the most promising methods for molecular genomics, but this technique is often associated with experimental complications and difficulties in the analysis. Moreover, the greatest part of genes displayed on an array is often not directly involved in the cellular process being studied. Recently, we proposed a data mining algorithm, based on the identification of genes involved in a given process, the calculation of interactions among them and their ranking according to number of interactions. Genes in the highest cluster are defined as "leader genes". These findings may lead to an ad hoc and therefore more significant experimentation. However, at present this complex process is performed manually. In this work, we present the general architecture of LeaderGene, an automated tool for ab-initio molecular genomics. Three different and independent parts: (1) Identification of gene list; (2) Calculation of weighted number of links; (3) Genes clustering. Initial inputs are provided by user; then, output of part 1 and part 2, respectively, become inputs of parts 2 and 3. The development of an user-friendly software capable to automatically compute leader genes in a given cellular system will allow further progresses in this field of molecular genomics