Inferring Time-Lagged Association Rules from Microarray Time-Series Data

GALLO, CRISTIAN ANDRÉS; CARBALLIDO, JESSICA ANDREA; PONZONI, IGNACIO

Cordoba

Congreso; 2do Congreso Argentino de Bioinformática y Biología Computacional; 2011

Asociación Argentina de Bioinformática y Biología Computacional

Gene regulatory networks play an important role in the progression of life phenomena such as cell cycling, developmental biology, aging, and the progressive and recurrent pathogenesis of complex diseases, among others. In this regard, the amount of genome-wide time series data is becoming increasingly available, providing the opportunity to reverse engineer the time-delayed gene regulatory networks that govern the majority of these molecular processes. The aim of this paper consists on the reconstruction of gene regulatory networks from multiple genome-wide microarray time series datasets given as input. In this sense, a new model-free algorithm called GRNCOP2 (Gene Regulatory Network inference by Combinatorial OPtimization 2) was developed. The method employs combinatorial optimization of gene profile classifiers inferring potential time-delay relationships between genes.  Additionally, it overcomes the limitations of the original version by means of several improvements. The proposed algorithm was applied to a time series data composed of twenty yeast genes that are highly relevant for the cell-cycle study, and the results were compared against several related approaches. In order to assess and contrast the performance of the methods, several well known data mining metrics were measured regarding three freely available databases of associations between genes. Table 1 shows the achievements of the algorithms with an accuracy of 0.75 for both simultaneous and time-delayed rules. The outcomes have shown that GRNCOP2 outperforms the contrasted methods in several of the proposed metrics, and that the results are consistent with previous biological knowledge. Additionally, the ability of GRNCOP2 to perform genome-wide studies was assessed with several genome-wide time series datasets. In this regard, the proper functioning of the method was demonstrated with the realization of an ontological analysis, showing that the results are significant in biological terms since the genes of the discovered sub-networks were found to be highly related in statistical terms. A novel method for the inference of time-delayed gene regulatory networks from genome-wide time series datasets is proposed in this paper. The method was carefully validated with several publicly available data. The results have demonstrated that the improvements incorporated in this new version of the algorithm provide a usable and powerful model-free approach to dissecting high-order dynamic trends of gene-gene interactions, carefully validated with publicly available cell cycling datasets through the well recognized S. cerevisiae Genome Database.