ABSTRACT
Chromatin interactions create spatial proximity between distal regulatory elements and target genes in the genome, which has an important impact on gene expression, transcriptional regulation, and phenotypic traits. To date, several methods have been developed for predicting gene expression. However, existing methods do not take into consideration the impact of chromatin interactions on target gene expression, thus potentially reduces the accuracy of gene expression prediction and mining of important regulatory elements. In this study, a highly accurate deep learning-based gene expression prediction model (DeepCBA) based on maize chromatin interaction data was developed. Compared with existing models, DeepCBA exhibits higher accuracy in expression classification and expression value prediction. The average Pearson correlation coefficients (PCC) for predicting gene expression using gene promoter proximal interactions, proximal-distal interactions, and proximal and distal interactions were 0.818, 0.625, and 0.929, respectively, representing an increase of 0.357, 0.16, and 0.469 over the PCC of traditional methods that only use gene proximal sequences. Some important motifs were identified through DeepCBA and were found to be enriched in open chromatin regions and expression quantitative trait loci (eQTL) and have the molecular characteristic of tissue specificity. Importantly, the experimental results of maize flowering-related gene ZmRap2.7 and tillering-related gene ZmTb1 demonstrate the feasibility of DeepCBA in exploring regulatory elements that affect gene expression. Moreover, the promoter editing and verification of two reported genes (ZmCLE7, ZmVTE4) demonstrated new insights of DeepCBA in precise designing of gene expression and even future intelligent breeding. DeepCBA is available at http://www.deepcba.com/ or http://124.220.197.196/.
