Abstract
Metagenomics is a powerful tool for characterizing viral composition within environmental samples, but sample and molecular processing steps can bias the estimation of viral community structure. The objective of this study is to understand the inherent variability introduced when conducting viral metagenomic analyses of wastewater and provide a bioinformatic strategy to accurately analyze sequences for viral community analyses. A standard approach using a combination of ultrafiltration, membrane filtration, and DNase treatment, and multiple displacement amplification (MDA) produced DNA preparations without any bacterial derived genes. Results showed recoveries in wastewater matrix ranged between 60-100%. A bias towards small single stranded DNA (ssDNA; polyomavirus) virus types vs larger double stranded DNA (dsDNA; adenovirus) viruses was also observed with a total estimated recovery of small circular viruses to be as much as 173-fold higher. Notably, ssDNA abundance decreased with sample dilution while large dsDNA genomes (e.g., Caudovirales) initially increased in abundance with dilution before gradually decreasing with further dilution in wastewater samples. The present study revealed the inherent biases associated with different components of viral metagenomic methods applied to wastewater. Overall, these results provide a well-characterized approach for effectively conducting viral metagenomics analysis of wastewater and reveal that dilution can effectively mitigate MDA bias.