The prediction method for folding boundaries of Escherichia coli chromosome based on multi-source information

The circular Escherichia coli chromosome has been suggested to form a partition of the genome where it folds into a collection of sequentially consecutive domains. This paper presents a prediction method to give the location of the folding boundaries. Because genes involved in one domain tend to be co-expressed, the information of activation and co-expression level is extracted from multi-source data especially. The activation level of genes is described by frequencies of attending the pathways in different database, while Intensity Co-expression is defined for co-expression level. The two types of data are processed by wavelet analysis and are used to improve the optimization model for prediction. The objective function of this model is to minimize both the total number of unfoldings of these domains in the chromosome and total frequencies of folding boundaries. The multi-source information is also designed for the experimental verification. The experiment results show the good effect of the prediction method in Computational Biology.