LAUSR

Just over a quarter century ago, two teams of microbial ecologists observed an unusually large number of ‘viruslike particles’, or VLPs, in lake and marine environments – thousands to millions of times higher than previous findings (Bergh et al., 1989; Proctor and Fuhrman, 1990). Subsequent observations confirmed this central finding and demonstrated that the majority of these VLPs were viruses resulting from the (largely) lysis of microbial cells. It follows then that viruses could have the potential to significantly influence the structure and function of natural ecosystems. Yet, given technical barriers to measurement, the viral ecologists – whose seminal work launched the field – had relatively few observations upon which to base quantitative estimates of viral effects. Instead, early efforts required extrapolation from a few observations of density, diversity and induced mortality to establish global estimates of impact (Suttle, 2005). With time, these extrapolations became paradigms (Breitbart, 2012). But now, nearly three decades after these initial findings, the viral ecology toolkit is transformed (Brum and Sullivan, 2015) and its paradigms are ripe for re-assessment. Looking into our Crystal Ball reveals that by 2020, the viral ecology field will have shifted from predominantly ‘specialists’ to more broadly include ‘non-specialists’. This will occur both collaboratively, but also through long-standing protocol development efforts – enabling measurements of viruses and their influence to be routinely integrated into all studies of environmental microbiology. The confluence of advances in sequencing technology and community-supported bioinformatics platforms will make it feasible to more broadly measure microbial ecosystem properties. Engagement with models and modelers will help disentangle the ecologically relevant signals to better understand how microbial viruses impact ecosystems. The blueprint for the former is already playing out in the democratization of virus ecology specific community tools and database development (Wommack et al., 2011; Roux et al., 2014; Bolduc et al., 2016; Kindler et al., 2016). Similarly, collective efforts to quantitatively model virus–microbe interactions at scales from molecules to ecosystems have provided the foundation for tighter integration between empiricists and modelers (Weitz, 2015). With this broader interest in the role viruses play in diverse microbial ecosystems and expanded resources from which to study these impacts, our Crystal Ball reveals the revision of major paradigms in viral ecology. The first paradigm to be revised is that viruses are 10-fold greater in abundance than microbes. Already, reanalysis of decades of virus and microbial abundance data find that virus–host ratios vary widely and are not well-described by a 10:1 ratio (Knowles et al., 2016; Parikka et al., 2016; Wigington et al., 2016). The interpretation, cause and consequence of this variability remains uncertain. Our Crystal Ball suggests that the field will move beyond enumerating ‘total viruses’ by developing and leveraging new technologies to resolve lineagespecific virus–host interactions at the community level. For double-stranded DNA phage of microbes at least, genomics is already being embraced as the foundation for viral taxonomy (Simmonds et al., 2016) and elucidating viral populations in natural samples (Brum and Sullivan, 2015), as well as connecting viruses and hosts at community scales (Edwards et al., 2016; Roux et al., 2016). The picture is far less simple for giant dsDNA viruses (Wilhelm et al., 2016), as well as ssDNA and RNA viruses, and our Crystal Ball remains cloudy here. Where possible, however, such lineage-specific virus– host measurements will (i) help elucidate mechanisms driving total virus-to-microbe-ratios, (ii) present microbial ecologists the opportunity to study viruses specific to focal microbes and (iii) inspire collaborations between theorists and empiricists that will guide experiments towards sampling schemes that will better inform ecosystem models. The second paradigm to be revised is that viruses are ‘the most diverse biological entities’ on the planet. This paradigm has its roots in back-of-the-envelope calculations that the global ‘virosphere’ contains some 2 billion proteins – a number extrapolated from a handful of phage genomes (Rohwer, 2003). However, viruses may not be the most diverse biological entities at either the level of proteins or organisms. First, it seems that many proteins must repeatedly occur within virus genomes as