Identification of recurring inter and intrahost viral SNVs among Dengue clinical sera through metagenomic sequencing

Dengue is one of the most prevalent arboviral disease in humans causing an estimated 50-100 million yearly infections globally every year, with a continuous increase in the number of people at risk due to geographical... [ view full abstract ]

Dengue is one of the most prevalent arboviral disease in humans causing an estimated 50-100 million yearly infections globally every year, with a continuous increase in the number of people at risk due to geographical expansion of the disease. Dengue is caused by an RNA virus with known high mutation rates resulting to a genetically diverse inter and intrahost viral population. These variations are collectively known as haplotypes/quasispecies and believed to have roles in viral fitness, immune escape, and pathogenesis. Studies of dengue genomic variations on actual clinical samples have been hampered by very low viral frequency of the minority variants,rendering them undetectable by PCR and regular Sanger sequencing. Moreover, classical approaches in analyzing dengue variations entail clonal sequencing or targeted resequencing, which introduces PCR and culture biases on the detection of true low frequency variants. To minimize this bias, we have performed a metagenomics approach directly on acute-phase dengue-confirmed clinical sera. This method may provide us a near actual genomic profile of the virus from a sick patient. Deep sequencing was done on eight PCR-confirmed DENV-2 sera using Illumina MiSeq platform. Genome assembly by read-mapping was done using Bowtie2 and single nucleotide variants (SNVs) were identified using the LoFreq program. To validate the detected SNVs, two more variant analyzers were used – VirVarSeq and ViQuaS. Only the variants detected by all three tools, with at least 5% frequency are accepted as true SNVs. Complete coding region (Capsid to NS5gene) of the DENV-2 genome was successfully assembled on all eight samples with a coverage depth of about 700X-3900X and about 59000-564000 mapped viral reads. Analysis of the sequences initially showed that at least two sets of intrahost SNVs are consistent with identified SNVs across other samples. Genomic regions containing these recurring SNVs include the capsid-prM, envelope and nonstructural 5 (NS5) genes. Intrahost SNVs at nucleotide positions 945 and 977 in the envelope gene are present on five of the samples as detected by all three variant calling tools. These sites are seen within the domain I and domain II regions of the envelope sequence,which are known to form epitopes that bind human neutralizing antibodies. The evidence of high variability on these sites as well as on the entire envelope gene is suggestive of the viral mechanism to adapt to high selective pressures. The study generally illustrates the capability of obtaining actual viral genomic profile of acute-phase dengue clinical sera using a metagenomic deep sequencing approach without performing the pre-processing methods such as whole viral genome amplification and clonal sequencing. Specifically, identification of recurring variable sites within the dengue genome involves probable determination of SNVs that may have roles in viral persistence, fitness, immune escape and pathogenesis.