Reversible Jump Markov Chain Monte Carlo Method For Local Graph Alignment

Biological networks are usually very complex. How to search the network motifs from a complex network is the problem which has been concerned in biology. In the issue of searching network motifs, it is an essential matter that how to determine k, the number of sub-graphs participating graph alignment. RJ-MCMC method can be exactly used to deal with the exponential mixed model which has unknown number components.This article consists of three parts. The basic concept of local graph alignment is introduced and several statistics of local graph alignment are established in the first part. In the next part, a hierarchical model based on the mixed Bayesian model is built and prior information of some variables is proposed. In the final part, the six steps of reverse jump movement are given in detail on the basis of the mixed exponential model obtained in the first section. The RJ-MCMC method is used to search for network motifs, the theoretical basis and implementation steps of determining the number of sub-graphs participating local graph alignment and motif with RJ-MCMC are given simultaneously.