Viruses are presumably the most abundant biological entities on the planet, with the total number of virus particles exceeding by 10 times the total number of cells (25). Many viruses have a relatively small genome encoding for a few proteins: one of the smallest being the circovirus with a 1.7-kb genome coding only two proteins (11). Despite their apparent simplicity, viral biochemistry and replication mechanisms are more varied than those seen in the entire bacterial, plant and animal kingdoms (15, 19). Nearly every possible method for encoding information in nucleic acid is exploited by viruses, from single-stranded DNA to double-stranded RNA. Each of the 83 virus families has a different replication strategy which calls for unique proteins and unique enzymes (19). For example the replication cycles of Human herpesvirus 1 (HHV-1) and Ebolavirus (EBOV) have nothing in common (Figure 1). The dsDNA HHV-1 genome encodes 73 proteins and replicates in the host nucleus where new viral genomes are encapsidated before budding through the endoplasmic reticulum and then into vesicles that will release the virion out of the cell (6,26). The EBOV ssRNA genome encodes eight proteins, replicates in the host cell cytoplasm using its own RNA-dependent RNA polymerase complex and buds directly at the plasma membrane (2). These two disparate replication cycles only illustrate the tremendous variety of viral molecular biology. As a result, it is crucial to have a clear vision of a specific virus’ biology in order to understand its genome and protein functions. Yet this information is hardly available outside academic books.

Unlike Luca for cellular organism (14), there is no presumed common ancestor for viruses (12). Therefore current virus classification comprises seven independent classes, according to the Baltimore system (4). This classification is based on the nature of the nucleic acids in the virion particle: dsDNA, ssDNA, dsRNA, ss(+)RNA, ss(−)RNA, ssRNA(RT) or ssDNA(RT).

Virus host ranges can be quite narrow, e.g. the human hepatitis B virus which is strictly restricted to Human, or very large e.g. the rabies virus which seems to be able to successfully infect any mammal. Knowing the host tropism is essential to understanding the viral molecular biology. For example a dsDNA viral genome are transcribed differently in a bacteria or in a eukaryote. Moreover virus host range has a dramatic importance for public health, as illustrated by zoonosis like SARS, Ebola or Influenza that are caused by viruses able to mutate and cross hosts barriers, thus threatening the human population. For all these reasons the display of virus host tropism is highlighted in ViralZone. The hosts are indicated by a colour code for each virus genera on the taxonomy pages (Figure 2B). ViralZone display of hosts is restricted to the natural reservoirs. Vectors hosts, dead-end or laboratory hosts are not described here except if a human host/cell-line is involved.

Virus fact sheets provide concise and specific information on molecular biology, taxonomy, hosts, and epidemiological data (Figure 2C). The first tab: ‘General’ describes molecular biology, virion and genome organization, followed by a step-by-step description of the viral cell infection cycle. The database section allows easy access to NCBI nucleotide and UniProtKB protein entries, as well as to specific virology databases such as VIPR (http://www.viprbrc.org) VIPERdb (7,23), Descriptions of Plant Viruses (1) and VBRC (www.vbrc.org, virology.ca). Host and cell tropism are generally indicated, but the latter might be absent since this kind of data can be unknown, or difficult to access in the literature. Cell receptor(s) for virus entry are also listed and links are provided to relevant publication(s). Finally epidemiological data briefly describe associated diseases and virus transmission as well as vaccines (if available) or antiviral drugs effective against this virus.

Virions are very diverse in shape and structure; they can be enveloped with one to several lipid bilayers or naked, and the genetic material can be protected by one, two or even three capsids showing helical or icosahedric symmetry and whose size is often related to genome length: from 17nm (Porcine circovirus, 1.7kb) to 400nm in diameter (Mimivirus, 1200kb). Virion pictures and diagrams can be found in Virology books or publications, but often with heterogeneous quality, colours and resolution. We created 160 original virion diagrams for ViralZone covering all known viral families and genera described to date. All the figures share the same concept and resolution, with defined colours for each part of the viral particle. Virions with icosahedric capsid symmetry are represented first as seen from a cross-section, then with 3D-like picture showing precise capsid architecture (Figure 3). Structural proteins are coloured in the same way in virion and genome pictures.

Virus genomes are relatively small, mostly <50kb, and are therefore easy and relatively inexpensive to sequence. This has resulted in an exponential increase in the number of new virus isolates deposited in the sequence databases. Of the 851503 viral protein entries in UniProtKB, the species with the greatest number of open reading frames (ORF) deposited in UniProtKB is HIV-1 with 313532 different ORFs, while the human proteome only accounts for 77225 entries (Table 2) (UniProt release 15.12).

Swiss Federal Government through the Federal Office of Education; Science grant Swiss Institute of Bioinformatics (SIB) (www.isb-sib.ch). Funding for open access charge: Swiss Institute of Bioinformatics.