A multivariate data analytics approach revealed no footprint of APOBEC3 proteins in hepatitis B virus genome

The human genome encodes for a family of editing enzymes known as APOBEC3 (apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like3). Several member of this family such as APOBEC3G, APOBEC3F and APOBEC3H haplotype II are active against viruses such as HIV. These enzymes induce C-to-U mutations in the negative strand of viral genomes, which leads to multiple G-to-A changes, referred to as ‘hypermutation’. Mutation by these enzymes are sequence context dependent in HIV genome, in which APOBEC3G preferentially mutates G within GG, TGG and TGGG content while other members mutate G within GA, TGA and TGAA. However, the same context between these enzymes and HBV virus have never been explored. In this study, we aim to identify mutational footprint of APOBEC3 enzymes in HBV genome. To do this, we perform a multivariable data analytics technique to investigate the motif preference and possible sequence hierarchy of mutation by APOBEC3 enzymes in a large number of full genome HBV sequences from a wide range of naturally infected patients. This enables us to differentiate between normal and hypermutated sequences in terms of the representation of mono- to tetra-nucleotide motifs and allows us to identify motifs associated with hypermutation by different APOBEC3 enzymes in the HBV genomes. Our analyses reveal that either APOPEC3 enzymes are not active against HBV virus or inducing G-to-A mutations by these enzymes are not sequence context dependent in HBV genome.