AUTHOR=Li Jing , Tian Fengjuan , Zhang Sen , Liu Shun-Shuai , Kang Xiao-Ping , Li Ya-Dan , Wei Jun-Qing , Lin Wei , Lei Zhongyi , Feng Ye , Jiang Jia-Fu , Jiang Tao , Tong Yigang 

TITLE=Genomic representation predicts an asymptotic host adaptation of bat coronaviruses using deep learning

JOURNAL=Frontiers in Microbiology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2023.1157608

DOI=10.3389/fmicb.2023.1157608

ISSN=1664-302X

ABSTRACT=Coronaviruses (CoVs) naturally reserve in bats, occasionally cause infection and transmission in human being and other mammals. Our study aims to build a deep learning (DL) method to predict the adaptation of bat CoVs to other mammals. CoV genome was represented with a method of dinucleotide composition representation (DCR) for viral two main genes of ORF1ab or Spike. DCR features were firstly analyzed for the distribution among adaptive hosts, and then were trained with a DL classifier of Convolutional Neural Networks (CNN) to predict the adaptation of bat CoVs. Results demonstrated inter-host separation and intra-host clustering of DCR-represented CoVs for six host types of Artiodactyla, Carnivora, Chiroptera, Primates, Rodents/Lagomorpha and Suiformes. The DCR-based CNN with five host labels (without Chiroptera) predicted a dominant adaptation of bat CoVs to Artiodactyla hosts, then to Carnivora and Rodents/Lagomorpha mammals, and then to Primates. Moreover, a linear-like and asymptotic adaptation of all CoVs (except from Suiformes) from Artiodactyla, to Carnivora and Rodents/Lagomorpha, and then to Primates, indicating an asymptotic bats-other mammals-human adaptation. In summary, genomic dinucleotides represented as DCR indicate a host-specific separation and clustering predict a linear-like and asymptotic adaptation shift of bat CoVs from other mammals to human beings, via deep learning.