Genetic diversity and connectivity of the Ostreid herpesvirus 1 populations in France: a first attempt to phylogeographic inference for a marine mollusc disease

Abstract

The genetic diversity of viral populations is a key to understanding ther phylogeographic and dissemination history of viruses, but studying the diversity of whole genomes from natural populations remains a challenge. Molecular ecology approaches are commonly used for RNA viruses harboring small genomes, but have only rarely been applied to DNA viruses with large genomes. Here, we used the Pacific oyster mortality syndrome (POMS, a disease that affects oyster farms around the world) as a model to study the genetic diversity of its causative agent, the Ostreid herpesvirus 1 (OsHV-1) in the three main French oyster-farming areas. Using ultra-deep sequencing on individual moribund oysters and new bioinformatics methodology, we de novo assembled 21 OsHV-1 genomes. Combining whole-genome comparisons with phylogenetic analysis and quantification of major and minor variants, we assessed the connectivity of OsHV-1 viral populations between the three oyster-farming areas. Our results suggest that the Marennes-Oleron Bay represents the main source of OsHV-1 diversity, from where the virus has dispersed to other farming areas, a scenario consistent with current practices of oyster transfers in France. Here, we demonstrate that molecular ecology approaches can be applied to large-genome viruses to determine the extent of their genetic diversity and better understand the spread of viral populations in natural environments.

Evaluation of tangential flow filtration coupled to long-read sequencing for ostreid herpesvirus type 1 genome assembly

“…This is supported by the tremendous increase of the signal (OsHV-1 reads) to noise (oyster reads) ratio obtained in the sequencing data. Indeed, more than 60 % of the sequenced reads belonged to OsHV-1 whereas, in previous work, the ratio obtained from infected tissues ranged from 0,01% to 13,21% [31, 34]. While other virus enrichment methods like PCR or capture-based approaches can reach up to 100 % of viral data, the TFF approach does not suffer from amplification or hybridization bias caused by sequence variation over time due to genomic divergence [97].…”

“…The de novo assembly of the long-read pipeline produced one large contig that has been manually curated to remove one extra duplicated region. Because of uncertainty with the short-read assembly to position duplicated regions relative to unique region, we have depicted the Illumina assembly as a Non-Redundant (NR) genome, with only one copy of each R L and R S , as frequently done in the literature [34,77,78,99]. In contrast, the long-read assembly resulted in a genome with both inverted repeats with the help of a minimal manual edit.…”

“…In order to determine the relationships of the OsHV-1 genomes from this study with those known so far, we constructed a phylogenetic tree from the alignment of the two genomes we generated in the present study (viruses collected in 2016), seven OsHV-1 genomes available in public databases (viruses collected before 2016) and four representatives of the 21 genomes acquired by Delmotte-Pelletier et al [34] (viruses collected in 2018). This analysis confirms that the OsHV-1 virus sampled in Thau lagoon (France) in 2016 belongs to µVar clade and that genomes reconstructed from the same sample with either the Nanopore or the Illumina sequencing data are really close to each other (Fig.…”

“…The identified OsHV-1 contigs were then manually scaffolded in Geneious Prime (version 2020.2.3) to reconstruct the common virus architecture (U L -IR L -X-IR S -U S ) [7]. To overcome the problem of inverted repeated regions that cannot be confidently positioned with short-read data, we constructed non-redundant genomes (NR genomes) that contain only one copy of each repeated region (U L -IR L -X-IR S -U S ) an approach commonly used for Herpesvirus genome analyses [34,77,78]. Quality filtered reads were then aligned to the new constructed OsHV-1 NR genome to evaluate coverage uniformity.…”

“…To date, 50 fully assembled OsHV-1 genomes are available in public databases, those genomes have been sequenced and assembled using various protocol and sequencing technologies (Table S1, available in the online version of this article). Although recent genomes have been obtained from short-read mNGS [31][32][33][34], OsHV-1 complete genome assembly remains a complex process. Indeed, a short-read approach is ideal for the identification of small-scale variation like SNPs but this technology struggles to identify variations at a larger scale like structure variations (SV) or large inverted and repeated regions.…”