Identification of viruses with the potential to infect human

Zhang, Zheng; Cai, Zena; Tan, Zhiying; Lu, Congyu; Zhang, Gaihua; Peng, Yousong

The virus has caused much mortality and morbidity to humans, and still posed a serious threat to the global public health. The virome with the human-infection potential is far from complete. Novel viruses have been discovered at an unprecedented pace as the rapid development of viral metagenomics. However, there is still a lack of a method for rapidly identifying the virus with the human-infection potential. This study built several machine learning models for discriminating the human-infecting viruses from other viruses based on the frequency of k-mers in the viral genomic sequences. The k-nearest neighbor (KNN) model could predict the human-infecting virus with an accuracy of over 90%. Even for the KNN models built on the contigs as short as 1kb, they performed comparably to those built on the viral genomes, suggesting that the models could be used to identify the human-infecting virus from the viral metagenomic sequences. This work could help for discovery of novel human-infecting virus in metagenomics studies.