Environmental viromes reveal global virosphere of deep-sea RNA viruses


 On the earth, the most abundant and diverse life forms are viruses. Both DNA viruses and RNA viruses play important roles in marine ecosystems via regulating biogeochemical cycles. However, the virome of marine RNA viruses has not been explored so far. Here, we established the global virome dataset of deep-sea RNA viruses that were purified from 133 sediment samples collected from typical deep-sea ecosystems (hydrothermal vents, cold seeps, ocean basins and mid-ocean ridges) of the Pacific Ocean, Indian Ocean and Atlantic Ocean. A total of 85,059 viral operational taxonomic units (vOTUs) were identified, of which only 1.72% were hitherto known, indicating that the deep-sea sediment is a repository of novel RNA viruses. The vOTUs were classified into 20 viral families, including prokaryotic (7.09%) and eukaryotic (65.81%) RNA viruses, of which Retroviridae was the most abundant. Furthermore, circular genome analysis revealed 1,463 deep-sea RNA viruses with complete genomes, the majority of which were unclassified. The distribution and differentiation of deep-sea RNA viruses were mainly influenced by the habitat. Taken together, the deep sea is a reservoir of novel RNA viruses, and our study is the first to analyze the composition and diversity of RNA viruses in deep-sea ecosystems.

Osedax japonicus, indicating that marine RNA viruses can infect animals apart from 92 microorganisms 13 . Nevertheless, the deep-sea RNA viruses are largely unexplored. 93 The deep sea is one of the largest and the least explored ecosystems on the 94 earth 14,15 . Characterized by high pressure, low temperature, low nutrient levels, lack 95 of sunlight and hypoxia, the deep sea is a challenge for inhabitants to survive 16,17 . 96 These extreme conditions put a selective pressure on the deep-sea organisms to adapt 97 to the environment 16 . In this study, we analyzed the global virome of RNA viruses 98 purified from 133 deep-sea sediment samples. The results showed that deep-sea RNA 99 viruses are abundant and highly diverse, and most are hitherto unclassified. In 100 addition, the distribution of these viruses is affected by the environmental factors.

102
The global virome of deep-sea RNA viruses 103 A total of 133 deep-sea sediment samples were collected during the 26th, 30th, 104 34th, 39th, 40th and 45th cruises of the geomicrobiology cruise of China, which 105 traveled more than 3,200,000 km from 2010 to 2018 ( Fig 1A and Table S1). The  Table   109 S1). The depths of sampling stations ranged from 1,100 to 6,105 m, and the average 110 6 depth was 3,545.3 m (Table S1). The environmental types of sampling stations 111 represented the typical environments of deep sea all over the world.

112
The viruses were purified from the sediments (Fig 1B), and RNAs extracted from 113 the purified virions had no DNA contamination (Fig S1), which excluded the presence  To minimize sequencing errors, we established 400-bp fragment libraries of viral 119 metagenomic DNAs obtained from the genomic RNAs. Each sample yielded more 120 than 3.5 Gb clean data (Table S1) (Fig 2A and Table S2).

135
Since most viral metagenomic sequences cannot be annotated and classified 18 , 136 taxonomic classification of RNA viruses based solely on metagenomics is challenging.

137
Therefore, we established a dataset of the genomic sequences of known RNA viruses 138 from public databases, which included 1,963,257 sequences. Only 1.72% vOTUs 139 matched the known sequences (Fig 2B and Table S3), indicating that the deep-sea 140 sediment is a repository of novel RNA viruses.

141
The vOTUs were taxonomically assigned to 20 families of RNA viruses ( Fig 2C   142 and Table S3). Retroviridae was the most abundant family that accounted for 40.07% 143 of all known vOTUs ( Fig 2C). Depending on the hosts, 7.09% of the classified 144 vOTUs were prokaryotic, 65.81% were eukaryotic, and 27.1% could not be assigned 145 to a host ( Fig 2D and Table S4). The CVGs were classified  Table S4). 155 Taken together, we established a dataset of 85,059 deep-sea RNA viruses, which is 156 considerably larger than the viral RefSeq database of NCBI that includes only 2,094 157 RNA viruses (Fig 2I). Our findings indicate a highly diverse community of RNA   Table S5). However, the abundance of the common RNA viruses 166 was less than <1% ( Fig 3B). In addition, we also detected vOTUs that were unique to 167 each ecosystem, while 8,098 vOTUs were common to all habitats ( Fig 3C and Table   168 S6), representing the core RNA viruses in deep-sea ecosystems. The relative 169 abundance of these core RNA viruses was also less than 1% (Fig 3D). The most    The viral diversity index was significantly different across the three oceans and 185 four habitats (p<0.05) (Fig 3G, 3H and Table S7). The Pacific Ocean had the most 186 diverse RNA viruses, whereas the lowest diversity was observed in the Indian Ocean 187 ( Fig 3G). Among the different habitats, the cold seeps and hydrothermal vents had the 188 highest and lowest diversity of RNA viruses, respectively ( Fig 3H). Thus, the viral                             To assemble the contigs, the reads were trimmed with Trimmomatic (version:0.33, 526 default parameters) 30 . Subsequently the reads of each sample were assembled to 527 contigs using metaSPAdes 3.12.0 31 .

528
To obtain the viral contigs, three kinds of softwares, including VirSorter 32 ,

529
VirFinder 33 and VIBRANT 34 , were used. The contigs whose length  1.0kb were