The human virome is vast and diverse, with approximately 1013 virus particles per individual. Virus populations are heterogenous and vary across different body sites. In response to the NIH Human Virome Program’s initiative to identify and catalog the human virome, we are optimizing methods to capture viral diversity across sample types. We prepared a mock viral community consisting of eight known viruses and examined whether the current methods capture and identify them. The mock community included E. coli phages T4, lambda, M13, and MS2, Pseudomonas virus phi6, an adenovirus associated vector (AAV), murine hepatitis virus (MHV), and vaccinia virus (VV). We spiked the mock community into different human sample types, including saliva, oropharyngeal (OP) wash, nasopharyngeal (NP) swab, and dental plaque, then passed the samples through different virus enrichment protocols for Illumina sequencing. We found that although multiple displacement amplification (Genomiphi) had great amplification

efficiency, it introduced huge biases, greatly favoring small circular single-stranded DNA viruses like M13. In contrast, Multiple Annealing And Looping Based Amplification Cycles (MALBAC) and Complete Whole Genome Transcriptome Amplification Kit 2 (WTA2) recovered viruses with a relative abundance more similar to the input mock community. Vaccinia passed through the virus enrichment protocol relatively well despite its large size particle size. Novel virus recovery was more robust from unspiked saliva (13.69%) compared to the recovery from OP wash (0.24%), NP swabs (3.82%) and dental plaque (5.25%). To determine the effect of DNA modifications on sequencing detection of viruses, we mixed known quantities of phage lambda with either phage T4 with glucosyl hydroxymethylated cytosine (T4 glc-hmC), T4 with hydroxymethylated cytosine (T4 hmC), or T4 with unmodified cytosine (T4 C), and then extracted nucleic acids for Illumina sequencing. Results showed that these modifications did not block sequence detection. To determine the optimal storage condition for viruses in human samples, we spiked the mock community into different samples resuspended in different storage buffers. Samples were stored at -80oC overnight and enriched for VLPs after one freeze-thaw before nucleic acids were extracted for Illumina sequencing. We found that viruses survived well, with no detectable contaminating viruses introduced in VTM or oral cocktail. In summary, our study demonstrates that virus enrichment methods can significantly influence virome profiling outcomes, and that spike-in studies can be used for optimization.