Q 1: What is vHoT DB?
A: vHoT is a database that searches for mRNA targets of viral microRNAs.

Q 2: How is vHoT DB different from similar tools?
A: vHoT DB allows inter-species analysis and can thus be used to predict cellular mRNA targets as well as viral mRNA targets of virus-derived microRNAs.

Q 3: What is the motivation for developing vHoT?
A: Since the identification of virus-derived microRNAs, it has become of interest to see the effects of viral microRNAs on virus itself and its host cells. As an obligate parasite organism, a virus inevitably requires host cells for replication and survival. Virally infected host cells possess two different genomes (host and viral genomes), and there exist complex interactions between microRNAs and mRNA targets in the infected cells. Cellular microRNAs may have effects on virus by affecting viral mRNA targets, and viral microRNAs may inhibit the expression of host mRNAs and virus mRNAs. In order to elucidate the function of microRNA of interest, its target mRNAs are often predicted with the aid of computational methods followed by biological validation. Although many computational approaches have been made to predict mRNA targets of microRNAs, they are limited to search for the interactions between microRNA and mRNA targets in vertebrates and some lower animals such as flies and worms, and the interactions are intra-species. Little has been done to predict the inter-species interactions and especially to find the host targets of vial microRNAs. To address this limitation of existing approaches, we propose vHoT DB that can predict host mRNAs targeted by virus-derived microRNAs.

Q 4: What kinds of target prediction tools is vHoT based upon?
A: vHoT currently utilizes most popular microRNA target prediction algorithms including TargetScan (Lewis et al., 2003), RNAhybrid (Rehmsmeier et al., 2004) and miRanda (Enright et al., 2003), DIANA microT (Maragkakis et al., 2009), PITA (Kertesz et al., 2007), and more tools will be added if they turn out to be appropriate.

Q 5: Where can I get more detailed information on viral microRNAs such as sequences, accession IDs, and related references?
A: The viral microRNA data used in vHoT were derived from the published miRNA sequences and annotations provided by miRBase. An easy way to retrieve detailed information on miRNA is to click on miRNA accession IDs from the search result page, which will direct you to the corresponding miRNA page in miRBase.

Q 6: Where can I get the accession ID of a target gene?
A: The mRNA accession IDs used in vHoT are from the NCBI (National Center for Biotechnology Information) system. Please go to NCBI for further information.

Q 7: In the input menu "Select species", what does the "Viruses" option mean?
A: The term "Viruses" means the viral genome from which the viral microRNA you select is derived.

Q 8: Is it possible to search for the target sites of user-specified genes instead of the entire genome?
A: Yes. Please enter a comma-separated list of mRNA accession IDs of interest into the "Target gene(s)" box located at the bottom of the "advanced options" menu. mRNA accession IDs can be obtained from the NCBI web site.

Q 9: Is it possible to predict the targets of miRNAs of any other organisms but virus?
A: No, vHoT was developed only for the published viral miRNAs. However, the target species can be human, mouse, cow, rat, rhesus monkey and the virus itself.

Q 10: Where can I find more information on the miRNAs used as input?
A: The viral microRNA data used in vHoT were derived from the published miRNA sequences and annotations provided by miRBase. An easy way to retrieve detailed information on miRNA is to click on miRNA accession IDs from the search result page, which will direct you to the corresponding miRNA page in miRBase.

Q 11: In the search result page, can I get some more detailed information on a specific prediction result?
A: Yes, just click any instance of "Detected" in the search page. (Note that you will not see this if your input query returns an empty result.) A new window with more detailed information will pop up. The number following "Detected" indicates the number of target sites within the candidate target gene.

Q 12: Can I get some information on viral miRNAs that are recently identified?
A: Yes, the miRNA data used by vHoT are based on the miRBase database which updates miRNA-related data on a regular basis (every 3-4 months). vHoT regularly checks the update status of miRBase and reflects the changes in order to keep the information as up-to-date as possible.

Q 13: What are the parameters that appear in the "advanced options" menu?
A: They are the input parameters of the five algorithms (TargetScan, miRanda, RNAhybrid, DIANA microT and PITA) used by vHoT. Each algorithm used in vHoT needs some input parameters, and the "advanced options" menu provides the user the opportunity to specify the algorithm parameters of individual algorithms. miRNAs are known to bind to the 3'UTR (untranslated region) of target mRNAs. Each algorithm quantitatively represents the interactions between miRNA and taget mRNAs by using their own methods and predicts candidate target mRNAs.

Q 14: Do different algorithms show different results?
A: Yes, in most cases. TargetScan, miRanda, RNAhybrid, DIANA microT and PITA have their own unique methodologies to predict mRNA-miRNA interactions. Thus, these methods may produce different outcomes, and their accuracy and search speed may also vary. For example, under the default input condition, the search query "mghv-miR-M1-1 against mouse genome" would yield 3456, 6019, 11679, 317 and 48 candidate targets for TargetScan, miRanda, RNAhybrid, DIANA microT and PITA, respectively.

Q 15: In the "advanced options" menu, what are the union and intersection options for?
A: As explained above, each of the algorithms employed by vHoT may yield different outcomes, and you may want to view the union or intersection of the five different test results. Select "Intersection" (this is the default option) if you want to view the search results commonly predicted by all five algorithms. If you want to see the list of entire target mRNAs predicted by at least one algorithm (i.e., any of TargetScan, miRanda, RNAhybrid, DIANA microT, and PITA), then select "Union."

Q 16: What is the accuracy of each algorithm?
A: It has been reported that TargetScan shows FPR (False Positive Rate) of 22% in mammals, miRanda represents FPR of 24-39% in flies, and RNAhybrid shows SNR of 2.9:1 (Min and Yoon, 2010).

Q 17: In the "Detailed information" page, what does the "|" or ":" symbol in UTR pairing represent?
A: The "|" symbol represents a base-pairing between A and U, or G and C, whereas the '.' symbol shows a wobble pair between G and U.

Q 18: In the "Detailed information" page associated with the miRanda algorithm, some nucleotides in UTR pairing are in lower cases, whereas some are not. What is the rule behind this?
A: The nucleotides involved in miRNA-mRNA interactions are represented by upper-case letters and unrelated nucleotides by lower-case letters.

Q 19: Some miRNAs are annotated with asterisks. What does this mean?
A: During typical miRNA biogenesis, one strand of a RNA duplex is preferentially selected for incorporation into RISC and is called the mature miRNA or the guided strand. However, the other strand, known as the passenger, carrier or miRNA* (miRNA star) strand, is generally degraded. For example, bovine herpesvirus 1 generates bhv1-miR-B6 and bhv1-miR-B6* which are derived from the same hairpin structure as shown below.
          a  ug    a   g  ga       ---   g
cgcggagggc ag  cccg cgc gg  acguggc   ggc g
|||||||||| ||  |||| ||| ||  |||||||   |||
gcgccucccg uc  gggc gcg cc  ugcgccg   ccg a
          g  cu    c   g  uc       cga   c

From this hairpin structure, the upper strand with 5'-aagugcccgacgcggggaacgug-3' is converted to bhv1-miR-B6, and the lower strand of 5'-cgucuccggcgccggguccuggc-3' becomes bhv1-miR-B6*.

Q 20: What are the data sources of vHoT?
A: The data listed in vHoT can be categorized into two: miRNA-related and mRNA-related data. The miRNA-related data are from the miRBase database, and the mRNA-related data are from the NCBI system.

Q 21: How do I know whether some predicted viral miRNA-host mRNA pairs are experimentally validated or not?
A: Some viral microRNAs have experimentally validated mRNA targets. If you click the Accession of a miRNA in the "Search results" page, you will be directed to the corresponding page of the miRBase database, where you can find references and information for further investigation. Additionally, you can refer to databases that list experimentally validated targets such as TarBase 5.0 and miRecords. As of November 25, 2010, two viral miRNA-target mRNA pairs are registered in TarBase, and none in miRecords.

Q 22: In the "advanced options" menu, what do the "MFE threshold" and the "total context score threshold" parameters mean, and how did you determine their default values?
A: First, the term MFE stands for Minimum Free Energy and represents the free energy of a predicted miRNA-target mRNA duplex in kcal/mol. In general, the lower the free energy is, the more stable the secondary structure of miRNA-mRNA duplex is. Second, the total context score threshold means the sum of the site-type contribution, the 3' paring contribution, the local AU base paring contribution, and the position contribution (see below for the definitions of these terms). If you set a MFE threshold and a score threshold, vHoT selects the miRNA-mRNA pairs whose MFEs are below the specified threshold and whose context scores are above the specified threshold, respectively. The default value was chosen by referring to the MFEs or the total context scores of the experimentally validated miRNA-mRNA duplexes.

Q 23: Where can I find the definitions of the TargetScan terms such as site-type contribution, 3' paring contribution and local AU contribution?
A: According to the TargetScan web site, the site-type contribution reflects the average contribution of each site type. A more negative score is associated with a more favorable site (Grimson et al., 2007). The 3' pairing contribution reflects consequential miRNA-target complementarity outside the seed region. A more negative score is associated with a more favorable site (Grimson et al., 2007). The local AU content reflects the transcript AU content 30 nt upstream and downstream of predicted site. A more negative score is associated with a more favorable site (Grimson et al., 2007). The position contribution reflects the distance to the nearest end of the annotated UTR of the target gene. A more negative score is associated with a more favorable site (Grimson et al., 2007). The context score is the sum of the contribution of these four features:

  • site-type contribution
  • 3' pairing contribution
  • local AU contribution
  • position contribution

  • The context score percentile rank is the percentage of sites for this miRNA with a less favorable context score.

    Q 24: In UTR paring, why does TargetScan show fewer base parings between miRNAs and mRNAs than RNAhybrid and miRanda?
    A: TargetScan predicts biological targets of miRNAs by searching for the presence of conserved 8mer and 7mer sites that match the seed region of each miRNA. That is, TargetScan considers the base pairing only within the seed region, thus producing fewer results than the other two algorithms.