Comparison of whole genome sequencing approaches for Capripox viruses

Direct whole genome sequencing of Capripox virus genomes from diagnostic samples is not always straightforward. Low viral content in a sample, biased sequencing and subsequent assembly and mapping methods may all influence the outcome. In this study we have tested and compared six next generation sequencing approaches on a homogenized skin sample from a bull infected with LSDV. We compared enrichment vs. non-enrichment strategies, different library preparation methods, short read Illumina sequencing with long read sequencing methods. We found that methods that use an unbound transposon during tagmentation produced unbalanced results and lower target read yield versus methods that use other approaches to the tagmentation step. We further find that the use of hybrid capture probes increased the number of target reads. The result of subsequent mapping and assembly steps are influenced by the choice of reference when using reference-based assembly approaches. When using a short read sequencing approach we advise to use a transposon free method or a method with bound transposons for DNA fragmentation. These methods outperform kits that employ free transposons for DNA fragmentation when targeting AT-rich genomes. When mapping the reads it is best to use a reference for assembly that is as closely related as possible to the sample under study. Mapping problems can be resolved by long read sequencing which we recommend for denovo whole genome sequencing. Pacific Bioscience based long read sequencing outperforms Oxford Nanopore sequencing because it is less error prone. The ONT approach used, displays the same bias as the transposon based approach from Illumina and is therefore less suitable when attempting (Capri)pox whole genome sequencing.