AUTHOR=Li Jing , Zhuo Linlin , Lian Xinze , Pan Shiyao , Xu Lei 

TITLE=DPB-NBFnet: Using neural Bellman-Ford networks to predict DNA-protein binding

JOURNAL=Frontiers in Pharmacology

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2022.1018294

DOI=10.3389/fphar.2022.1018294

ISSN=1663-9812

ABSTRACT=DNA, as hereditary materials, plays an essential role in micro-organism and almost all other organisms. And proteins are a vital composition and principal undertaker of microbe movement. Therefore, studying the bindings between DNA and proteins is of high significance from the micro-biological point of view. Also, the binding affinity prediction is beneficial for study on drug design. However, existing experimental methods for identifying DNA-protein bindings are highly costly and time-consuming. To solve such a problem, many deep learning methods including graph neural networks have been developed to predict DNA-protein interactions. Our work possesses the same motivation and we put the latest NBFnet (Neural Bellman-Ford Neural Networks) into use to build pair representations of DNA and protein to predict the existence of DNA-protein binding (DPB). NBFnet is a graph neural network model which utilized the Bellman-Ford algorithms to get pair representations and has been proved to have a state-of-the-art performance solving the link prediction problem. After building the pair representations, we designed a feed-forward neural network structure and get 2-D vector output as predicted value of positive or negative samples. We conducted our experiments on 100 datasets from ENCODE datasets. Our experiments indicated that DPB-NBFnet has shown its competitive performance compared with the baseline models. Apart from this, we have executed parameter tuning with different architectures to explore the structure of our framework.