Probabilistic consistency transformation for multiple alignment of biological networks

In this paper, we propose a probabilistic network alignment approach that maximizes the expected accuracy of the alignment. To this aim, we define a set of correspondence scores between each node pair of two networks using semi-Markov random walk. To increase the consistency of the alignment, we then update these scores by incorporating information from other networks. We employ the transformed scores into a greedy alignment process. Experiments reveal that proposed approach can enhance the alignment accuracy.