The Pacific Ocean virome (POV): a marine viral metagenomic dataset and associated protein clusters for quantitative viral ecology.

Hurwitz BL, Sullivan MB. · 2013

Read The Pacific Ocean virome (POV): a marine viral metagenomic dataset and associated protein clusters for quantitative viral ecology. by Hurwitz BL, Sullivan MB. (2013) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc3585363.txt.

Introduction

Bonnie Hurwitz and Matthew Sullivan assembled a Pacific Ocean virus data set in PLoS ONE in 2013. Ocean viruses, mostly killers of bacteria, steer nutrient cycles and gene flow, yet most viral genes read as unknown in databases, far more than for bacteria. Without an organized map, that dark sequence space stays unusable. Their project gathered virus samples across the Pacific to build one.

Thirty-two viral metagenomes came from four regions, spanning seasons, depths down to the deep open ocean, and distance from shore. Sorting the proteins into families produced about twenty-seven thousand high-confidence clusters, more than doubling what existed for viruses. Richness of these families fell from deep water to the surface, from winter to summer, and offshore in surface waters. The set gives ecologists a shared reference for counting and comparing ocean viruses quantitatively.

Catalog record

CREATORHurwitz BL, Sullivan MB.
YEAR2013
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC3585363

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The Pacific Ocean Virome (POV): A Marine Viral Metagenomic Dataset and Associated Protein Clusters for Quantitative Viral Ecology By Bonnie L Hurwitz, Matthew B Sullivan, Fabiano Thompson Published in: PLoS ONE DOI: 10.1371/journal.pone.0057355 ABSTRACT Abstract Bacteria and their viruses (phage) are fundamental drivers of many ecosystem processes including global biogeochemistry and horizontal gene transfer. While databases and resources for studying function in uncultured bacterial communities are relatively advanced, many fewer exist for their viral counterparts. The issue is largely technical in that the majority (often 90%) of viral sequences are functionally ‘unknown’ making viruses a virtually untapped resource of functional and physiological information. Here, we provide a community resource that organizes this unknown sequence space into 27 K high confidence protein clusters using 32 viral metagenomes from four biogeographic regions in the Pacific Ocean that vary by season, depth, and proximity to land, and include some of the first deep pelagic ocean viral metagenomes. These protein clusters more than double currently available viral protein clusters, including those from environmental datasets. Further, a protein cluster guided analysis of functional diversity revealed that richness decreased (i) from deep to surface waters, (ii) from winter to summer, (iii) and with distance from shore in surface waters only. These data provide a framework from which to draw on for future metadata-enabled functional inquiries of the vast viral unknown. INTRODUCTION Bacteria are fundamental to life on Earth and drive energy and nutrient cycles in natural systems [1]. In the surface oceans, viruses, thought to be predominantly phages, and are present at approximately 107 particles per milliliter of seawater often outnumbering their bacterial hosts fifteen to one [2]–[4]. These abundant phages play an integral role in the lifecycle, development and evolution of their diverse hosts (reviewed in [5]). This is particularly well documented in ocean cyanophages where natural abundances are high [6]–[9], and the eco-evolutionary virus-host dynamics involve manipulation of critical aspects of ocean life (e.g., photosynthesis, phosphate, nitrogen; reviewed in [10]). Despite this importance, our understanding of indigenous viral and microbial communities is compromised by both conceptual and technological challenges. Perhaps chief among them is that <1% of microbes in an environment can be routinely grown in the laboratory [11], [12]. To study this ‘unseen majority’ [13], microbial ecologists developed culture-independent techniques, first in the form of gene markers (e.g., small subunit ribosomal DNA, [14]) to quantify biodiversity and subsequently through metagenomics to study metabolic function and gene ecology [12], [15]–[17]. To this end, resources such as the Global Ocean Sampling (GOS) dataset [18] have been developed, and subsequently organized using sequence-based protein clusters (PCs) [19] to (i) group related proteins based on sequence similarity, (ii) map metadata associated with the samples to the PCs, and (iii) aid future inquires in genetic novelty by mapping to pre-defined PCs. The GOS dataset represents 7.7 million reads from 41 surface ocean microbial (0.1–0.8 µm size fraction) samples ranging 8,000 km along coastal Atlantic Ocean eastern United States waters to the Sargasso Sea, Gulf of Mexico, Caribbean Sea and a few sites in the South Pacific Ocean. This broad sampling has enabled GOS to greatly advance microbial ecology (849 citations as of 6 Sept 2012; Google Scholar) both through data mining studies and use of GOS to contextualize environmental research findings. While metagenomics was first applied to uncultured viral communities [20], two years in advance of microbes [15]–[17], there remains no such equivalent GOS-scale datasets (Table 1). Such a resource is missing because of limiting DNA yields from environmental viruses and the fact that viral metagenomic analytics are challenging. On the latter, viral metagenomics is more difficult than microbial metagenomics as (i) viral genome representation in public databases is more sparse (as of November 2012 there are 115 marine phage genomes in Genbank, Table S1), (ii) analytical tools paralleling those for microbes (e.g., MG-RAST [21] and IMG/M [22]) have not been available (but see VIROME [23] and MetaVir [24]) and (iii) the ‘unknown’ problem is even greater in viral metagenomes as often 90% of sequences lack similarity to anything in pre-existing protein databases (see Table 1 versus 65% for microbes [17]). Worse, the viral metagenomic datasets themselves often suffer either from being too small (e.g., LASL-based Sanger-sequenced metagenomes) or contain technical artifacts that are now known to render the metagenomes non-quantitative. To emphasize the latter, we highlight a commonly used marine viral dataset (363 citations as of 6 Sept 2012; Google Scholar) that includes 1.77 million metagenome sequences from 4 major ocean regions [25]. While this dataset was generated using the best available technologies and is the gold-standard for contextualizing viral ecological findings, it is now known to suffer from two major technical issues: (i) pooled samples (4 metagenomes from 184 viral assemblages collected over a decade from 68 sites) inhibit tracking of variability over space and time, and (ii) MDA-amplified material is now known to lead to non-quantitative metagenomes [26]. This latter point is particularly problematic for viruses as MDA biases are both stochastic [26] and systematic [27], [28] with the latter related to nucleic acid type and structure (i.e., a feature which varies across viral types). For these reasons, there are currently no marine viral metagenomic datasets appropriate for quantitative viral ecology. [Table: Summary of published marine, non-coral viral metagenomic datasets.] Here we introduce a large-scale, quantitative Pacific Ocean Virome (POV) dataset and ∼456 K associated PCs that organize the ‘known’ and ‘unknown’ sequence space for future comparative viral metagenomic study. The 6 million read dataset is derived from 32 temporally- and spatially-resolved viral assemblages, and represents the largest viral metagenomic sampling of the Pacific Ocean to date, including the first large-scale viral metagenomes from the deep pelagic ocean (but see [29], [30] and Table 1). The POV dataset represents a systematically collected, processed, and documented quantitative marine viral metagenomic resource [31] as follows. All thirty-two POV communities were concentrated using a new method that captures nearly all particles [32], purified using DNase digestion and CsCl buoyant density gradients to minimize contamination by non-viral DNA [33], and DNA extracted and linker-amplified to minimize quantitative and cloning biases in the resulting metagenomes [34]. DNA was then sequenced by Roche 454 Titanium technology. The metagenomes and the associated PCs provide a much-needed community resource to test hypotheses about environmental viruses, as GOS has done for microbial ecology. For these reasons, POV will likely become a foundational dataset for future comparative studies of virus genes and communities at the global ocean scale such as those derived from the recent Tara Oceans [35] and Malaspina [36] expeditions. RESULTS AND DISCUSSION The Pacific Ocean Samples The 32 POV source waters varied by depth, proximity to land, and season and were derived from four regions in the Pacific Ocean (Figure 1): Scripps Pier in San Diego, California (SIO), Monterey Bay, California (MBARI), near Vancouver Island in British Columbia (LineP), and the Great Barrier Reef in Australia (GBR). Samples, metadata, and metagenomic descriptive statistics are summarized in Table 2. [Figure: Sampling site map for the POV dataset.Thirty-two viral metagenomes represent discretely sampled and processed datasets that vary over time and space in the pelagic Pacific Ocean. (A) Overview of sampling sites, (B) GBR – Great Barrier Reef, Australia, near Dunk and Fitzroy Islands.

Read the complete text