Prediction of enhancer–promoter interactions using the cross-cell type information and domain adversarial neural network

Abstract

Background Enhancer–promoter interactions (EPIs) play key roles in transcriptional regulation and disease progression. Although several computational methods have been developed to predict such interactions, their performances are not satisfactory when training and testing data from different cell lines. Currently, it is still unclear what extent a across cell line prediction can be made based on sequence-level information. Results In this work, we present a novel Sequence-based method (called SEPT) to predict the enhancer–promoter interactions in new cell line by using the cross-cell information and Transfer learning. SEPT first learns the features of enhancer and promoter from DNA sequences with convolutional neural network (CNN), then designing the gradient reversal layer of transfer learning to reduce the cell line specific features meanwhile retaining the features associated with EPIs. When the locations of enhancers and promoters are provided in new cell line, SEPT can successfully recognize EPIs in this new cell line based on labeled data of other cell lines. The experiment results show that SEPT can effectively learn the latent import EPIs-related features between cell lines and achieves the best prediction performance in terms of AUC (the area under the receiver operating curves). Conclusions SEPT is an effective method for predicting the EPIs in new cell line. Domain adversarial architecture of transfer learning used in SEPT can learn the latent EPIs shared features among cell lines from all other existing labeled data. It can be expected that SEPT will be of interest to researchers concerned with biological interaction prediction.

A systematic study of motif pairs that may facilitate enhancer–promoter interactions

“…Moreover, EBF1, ZNF143, and RAD21 have a three-way interaction in GM12878 [ 56 ]. It is thus likely that the interaction of EBF1-ZNF143 may contribute to EP interactions [ 22 ].…”

“…With the transcription start sites (TSSs) defined in GENCODE, we defined 57,820 promoters, each of which was the genomic region from the upstream 1000 bps to the downstream of 100 bps the TSS of a GENCODE gene. An active promoter was then defined with these GENCODE promoters and the ENCODE RNA-seq data as previously [ 22 , 24 ]. In this way, every positive EP pair had its enhancer overlapping with one genomic region and its promoter overlapping with the other genomic region of a positive pair of genomic regions, and the distance between the active enhancer and the active promoter was within 2.5 kilobase pairs to 2 megabase pairs ( Supplementary S1 ).…”

“…Several studies pointed out a new venue to explore the characteristics of EP interactions, which suggested that the interaction of transcription factors (TFs) that bind an enhancer and TFs that bind a promoter of an EP pair may contribute to the interaction of this EP pair [ 2 , 12 , 19 , 22 , 27 ], [ 28 ], [ 29 ], [ 30 ]. For instance, it is well known that the TF and structural protein CTCF binds to a fraction of enhancers and promoters, which facilitates the physical interaction of enhancers and promoters in these EP pairs [ 31 ].…”

EPIsHilbert: Prediction of Enhancer-Promoter Interactions via Hilbert Curve Encoding and Transfer Learning

“…They obtained satisfactory prediction accuracy when train datasets were the same as test datasets, but performed worse across cell lines. Although EPIHC [ 23 ] and SEPT [ 24 ] both use different transfer learning approaches, their results are unsatisfactory.…”

A Systematic Study of Motif Pairs That May Facilitate Enhancer-Promoter Interactions

“…Several studies pointed out a new venue to explore the characteristics of EP interactions, which suggested that the interaction of transcription factors (TFs) that bind an enhancer and TFs that bind a promoter of an EP pair may contribute to the interaction of this EP pair [2,12,20,[27][28][29][30][31]. For instance, it is well known that the TF and structural protein CTCF binds to a fraction of enhancers and promoters, which facilitates the physical interaction of enhancers and promoters in these EP pairs [32].…”