Abstract: Aiming at the problems of insufficient data and inaccurate results of constructing gene regulatory network algorithm based on single dataset, we proposed a data source fusion algorithm based on availability and responsibility (AP). The gene expression data, protein interaction data and motif data were analyzed and constructed gene regulatory network by controlling and controlled twoway flowing information transmission respectively, and compared with AUC values of three developed models: ReMoDiscovery, CLR and C3Net in yeast genomewide network construction. The comparison results show that the algorithm has higher execution efficiency and stronger convergence in the construction of gene regulatory network algorithm.

Key words: AP algorithm, data source fusion, gene regulatory network, genomewide association