Viromes outperform total metagenomes in revealing the spatiotemporal patterns of agricultural soil viral communities

Soil viruses are abundant and influence biogeochemical cycles by controlling microbial populations, redirecting metabolism, and mediating gene transfer. Their ecological roles in soil are still poorly understood, despite their potential importance in carbon and nutrient cycling. Viral abundances are linked to host community composition, and agricultural practices (like rhizosphere processes and soil amendments) can alter both microbial and viral diversity. Environmental factors like temperature, pH, and nutrient status may also impact viral communities. Metagenomic approaches are crucial for studying viral diversity due to the absence of a universal marker gene. Total metagenomes, while convenient, can be dominated by bacterial and eukaryotic sequences, obscuring the less abundant viral signal. Viromes, generated by size-fractionation, enrich viral sequences and offer a more complete picture of viral diversity. However, soil viromics has been challenging due to low DNA extraction yields. Recent improvements in DNA extraction now allow for the efficient generation of soil viromes from manageable soil quantities. This study used both total metagenomes and viromes to characterize soil dsDNA viral communities in an agricultural tomato field, comparing their effectiveness in recovering viral sequences and exploring viral and microbial ecological patterns.

The introduction section extensively reviews the existing literature on soil viruses and their ecological roles, highlighting the challenges in studying them using traditional metagenomic techniques. It emphasizes the limitations of total metagenomes in capturing the full diversity of soil viral communities, particularly the rare biosphere. The advantages of using viromes, while acknowledging the historical challenges associated with their preparation from soil samples, are clearly outlined. The review establishes the need for a comparative study using both methods to assess their respective capabilities in revealing the complexity and dynamics of soil viral communities.

Soil samples were collected from a tomato field at pre-planting and harvesting stages. Eight plots were sampled, each with different biochar treatments and nitrogen fertilization regimes. Soil chemistry and moisture content were measured. For viromics, viral size fractionation was performed using 0.22 µm filtration and ultracentrifugation to concentrate purified virus-like particles. DNase treatment removed free DNA before extraction with the PowerSoil kit. Total DNA for metagenomics was extracted from 0.5 g of soil. Libraries were constructed using different kits (DNA Hyper Prep and Nextera DNA Flex) for different time points. Paired-end sequencing (150 bp) was performed using Illumina HiSeq 4000. Quality filtering was performed with Trimmomatic and BBDUK. De novo assembly was done with MEGAHIT, and clustering with PSI-CD-HIT. VirSorter and DeepVirFinder identified viral contigs. vConTACT2 provided taxonomic classifications. Read mapping was performed with BBMap, and vOTU coverage tables generated with BamM. 16S rRNA gene fragments were identified using SortMeRNA and the RDP classifier. K-mer profiling was done with sourmash. Statistical analyses were done using R.

Viromes significantly outperformed total metagenomes in recovering viral sequences. Viromes showed a significant depletion of bacterial and archaeal sequences (0.006% of reads vs 0.042% in total metagenomes). Total metagenomes exhibited higher sequence complexity and lower assembly quality. Viromes yielded 800 Mbp of assembled sequences vs 65 Mbp in total metagenomes. 52.4% of virome contigs were identified as viral vs 2.2% in total metagenomes. 2961 vOTUs were detected, with 2684 exclusively from viromes and only 3 from total metagenomes alone. Viromes captured a more complete representation of the rare virosphere. Highly abundant vOTUs in viromes had high occupancy across samples, while rare vOTUs were found in fewer samples. Total metagenomes showed a sparse recovery of viral diversity. Abundance-based rank of vOTUs was not consistently preserved between total metagenomes and viromes. Both viral and microbial communities showed significant spatiotemporal patterns. Viral communities alone exhibited strong spatial structure.

The study demonstrates the superiority of viromes over total metagenomes for characterizing soil viral communities, particularly in revealing the less abundant and more diverse rare virosphere. The findings highlight the limitations of relying solely on total metagenomes for comprehensive viral community analysis. The observed spatiotemporal patterns suggest coupled responses of viral and microbial communities to environmental factors and agricultural practices. However, differences in spatial distribution suggest potential decoupling of the processes driving viral and microbial community dynamics. This decoupling may result from differences in viral dispersal mechanisms, decay rates, and/or sensitivities to soil heterogeneity.

Viromes are a superior approach for characterizing the diversity and dynamics of soil viral communities. The study reveals the importance of considering both spatial and temporal factors in understanding viral ecology. Future research could explore the specific mechanisms driving the observed decoupling of viral and microbial community distributions and examine the functional roles of the identified viral populations in soil biogeochemical cycles.

The study used a limited number of samples and a single agricultural field. Differences in library construction methods between April and August samples could potentially introduce biases. Differential adsorption of viruses to the soil matrix might affect virion recovery and relative abundance estimates.