An intelligent agents architecture for DNA-microarray data integration

DNA microarray is a highly-parallel methodology to screen genome samples for the expression of particular DNA sequences [1]. To gather information about gene expression is a much more difficult task than DNA sequence, because gene expression is a dynamic process dependent on many factors. The knowledge of the ORFs (Open Reading Framework) for a certain organism opens the possibility to globally study the expression of these genes under chosen circumstances. By using DNA-microarray methodology, the investigator gets a global picture of the genes expression under study. Since experimental data are disseminated in a myriad of different databases that are duplicated in several repositories, the experimental data integration and clustering imply the exploration of a sites network from which the investigator can integrate (select and cluster) data of interest. The authors propose to adopt an intelligent agent architecture as communication and coordination tool among distributed sites. In this paper an intelligent agent architecture will be applied to the case of the expression of yeast genome, where several microarray hybridization experimental datasets are available [9]. Introduction DNA microarray is a highly-parallel methodology to screen genome samples for the expression of particular DNA sequences [1]. In 1996 it has been successfully used for the yeast genome mapping [2]. Lately, several diagnostic applications have been developed and DNA-array technique is believed to be a promising tool in clinical molecular biology for the years to come [3]. Several governative labs and universities together with increasing private facilities had started to build their own databases holding their DNA-array experimental data [4]. In this framework, the massive and distributed amount of DNA-array data is demanding for reliable tools to communicate and mine the stored information; in other words, it would be preferable to have a standardized format for storing DNA-array experimental data and to have on handle intelligent agents to search for user queries in the distributed DNA-array datasets ensembles. European Molecular Biology Lab (EMBL) and the European Bioinformatic Institute (EBI) had recently submitted to OMG (Object Management Group) a standard format for microarray experimental data (MAML) while other governative labs had proposed similar but different formatting rules as GEML (proposed by a public-private coomunity, the GEML community [21]).or BSML (Bioinformatic Sequence Markup Language) proposed by Visualgenomic, Inc. USA [20]. Functional genomics Functional genomics is the science of understanding how the genome functions through controlling the expression of genes. At present, the main application of microarray hybridization experiments is the concurrent evaluation of the impact of different conditions (pathological states, environmental stress, etc) on the expression of multiple genes. In fact, the majority of states in health or disease is most likely controlled by hundreds of genes. Until recently, molecular biomedicine could study one or very few genes in parallel. With the advent of the high throughput micro-array technology it is now possible to observe the levels of activity in literally thousands of genes. Besides other necessary computational efforts regarding data clustering [5], mining, and model-based/model-free [6] functional classification [7], micro-array hybridization experiments require a propedeutic step where preprocessing tasks have to be pursued [8]. In fact, several micro-array experimental datasets could be made available from different labs (at the same time or consecutively) for the same pathology or gene expression system. So it would be preferable to manage and