Between- and within-host mutation of dengue virus

RNA viruses exhibit high mutation rates, leading to multiple haplotypes (also referred to as quasi-species) within infected hosts and have been described in both acute (e.g. Influenza, DENV) and chronic (e.g. HIV, HCV) infections. This study characterises the mutation hotspots in DENV genomes at both consensus and haplotype levels. Near full-length DENV genomes were sequenced using Oxford Nanopore Technology (ONT) from Sri Lankan dengue patients recruited between 2017 and 2020. Consensus sequences were made with Minimap-2, and haplotypes were reconstructed with Nano-Q, a tool designed for estimation of RNA virus haplotypes and their relative abundance. The genomic variability of DENV genomes was assessed using Shannon Entropy (SE). Codons undergoing diversifying selection were identified with three phylogenetics-based algorithms (FEL, MEME, FUBAR) implemented within the Datamonkey suite. Hotspots were defined as the high entropy codons that were also subject to diversifying selection. From 150 samples tested, both consensus and haplotype sequences were characterised in 90 samples (DENV1: 8, DENV2: 51, DENV3: 31). At the consensus level, the NS2A gene had the greatest number of sites with higher Shannon entropy when adjusted for gene length across all serotypes, while at the haplotype level the NS1 gene had the same. Overall, the haplotypes sequences revealed more sites with high mutability and codons under diversifying selection than those visible at consensus level. This provides proof-in-principle that in acute RNA viruses also have high mutability in haplotypes, which may be inapparent with a consensus-level analysis.