Higher-order partial least squares for predicting gene expression levels from chromatin states

Multiple publications have indicated that gene expression levels are strongly affected by chromatin mark combinations via at least two mechanisms, i.e., activation or repression. But their combinatorial patterns remain unresolved. To further understand the relationship between histone modifications and gene expression levels, here in this paper, we introduce a purely geometric higher-order representation, tensor (also called multidimensional array), which might contain more hidden information from chromatin states to predicting gene expression levels. The prediction models were learned from regions around upstream 10k base pairs and downstream 10k base pairs of the transcriptional start sites (TSSs) over three species (i.e., Human, Rhesus Macaque, and Chimpanzee) with five histone modifications (i.e., H3K4me1, H3K4me3, H3K27ac, H3K27me3, and Pol II). Experimental results demonstrate that the proposed method is more powerful for predicting gene expression levels than several commonly used methods. Specifically, our method improves the performance on both criteria, R and RMSE as high as 1.7% and 11%, respectively.