Interpretable detection of novel human viruses from genome sequencing data

Jakub M Bartoszewicz; Anja Seidel; Bernhard Y Renard

Within a globally interconnected and densely populated world, pathogens can spread more easily than they ever did before. As the recent outbreaks of Ebola and Zika viruses have shown, the risks posed even by these previously known agents remain unpredictable and their expansion hard to control (Calvignac-Spencer et al., 2014) . What is more, it is almost certain that more unknown pathogen species and strains are yet to be discovered, given their constant, extremely fast-paced evolution and unexplored biodiversity, as well as increasing human exposure (Vouga & Greub, 2016; Trappe et al., 2016) . Some of those novel pathogens may cause epidemics (similar to the SARS and MERS coronavirus outbreaks in 2002 and 2012) or even pandemics (e.g. SARS-CoV-2 and the "swine flu" H1N1/09 strain). Many have more than one host or vector, which makes assessing and predicting the risks even more difficult. For example, Ebola has its natural reservoir most likely in fruit bats (Leendertz et al., 2016) , but causes deadly epidemics in both humans and chimpanzees. As the state-of-the art approach for the open-view detection of pathogens is genome sequencing (Lecuit & Eloit, 2014; Calistri & Palù, 2015) , it is crucial to develop automated pipelines for characterizing the infectious potential of currently unidentifiable sequences.