Build:


Release:		Devel:

miRNAmeConverter

When working with a set of miRNAs, useful information includes

if a certain miRNA name exists in the miRBase database.
the miRNA sequence in a certain miRBase version.
if a given miRNA is still considered to be a miRNA.
the miRNA names from the most recent miRBase version
the respective miRNA name from a different miRBase version, for using other services, such as miRNA target databases.

The miRNAmeConverter R package has been developed for handling naming challenges of mature miRNAs.

In addition we have developed a web interface that enables one to use the translateMiRNAName function via web interface and is based on shiny miRNAmeConverter-web.

Introduction
The MiRNANameConverter class
Use Cases
Additional Information

Please support this work

If you make use of this package, please cite the following ‘OpenAccess’ publication to allow us to keep the package up to date:

Haunsberger SJ, Connolly NMC and Prehn JHM* (2016). “miRNAmeConverter: an R/Bioconductor package for translating mature miRNA names to different miRBase versions.” Bioinformatics. doi: 10.1093/bioinformatics/btw660.

Introduction

The miRNAmeConverter package delivers results in a fast and transparent way. The main functions

check for validity of mature miRNA names,
determine the most likely miRBase version of a given set of miRNAs and
translate mature miRNA names to different versions (incl. sequences).

The core function, translating miRNA names to different versions, is especially useful when dealing with miRNA tools other than miRBase. To retrieve targets from miRTarBase, for example, the miRNA name from version 20 is required, whereas the website miRecords only accepts version 17. The miRNAmeConverter can manage large sets of miRNA names and hence can be easily implemented into workflows.

The data set included in the package (miRNAs) is a collection of human miRNA names. It consists of valid miRNA names (some duplicates), incorrect names and features used as controls in experiments, such as the RNU44 as a house keeping gene for HT-qPCR assay plates from Applied Biosystems.

To load the package and gain access to the functions and sample dataset of the miRNAmeConverter package just run the following command:

> library(miRNAmeConverter)

miRNA name input values

According to the nomenclature used in the miRBase repository hsa-mir-29a is a stem-loop sequence name. If we substitute the ‘r’ by a capital ‘R’ it codes for the mature 3’ miRNA hsa-miR-29a (or hsa-miR-29a-5p in the current version). The miRNAs input value for the functions has to be in form of a character vector. Algorithms of the package are not case sensitive. This has the effect, that for example in the case hsa-mir-29a and hsa-miR-29a as input values both will be considered as valid mature miRNA names.

The MiRNANameConverter class

The MiRNANameConverter class is coded in S4 style. All functions offered by the miRNAmeConverter package are methods of that class. All methods can be displayed by

> ls("package:miRNAmeConverter")

#  [1] "assessVersion"           "checkMiRNAName"          "currentVersion"         
#  [4] "example.miRNAs"          "getMirbaseVersionsXandY" "MiRNANameConverter"     
#  [7] "nOrganisms"              "nTotalEntries"           "saveResults"            
# [10] "show"                    "translateMiRNAName"      "validOrganisms"         
# [13] "validVersions"

The slot names (attributes) of the class can be displayed by

> slotNames("MiRNANameConverter")
[1] ".dbconn"         ".currentVersion" ".validVersions"  ".nOrganisms"    
[5] ".nTotalEntries"  ".validOrganisms"

An instance of the MiRNANameConverter class can be created by calling the MiRNANameConverter function:

> MiRNANameConverter()
An object of class: MiRNANameConverter
- Most recent miRBase version provided in the package:[1] 22.1
- Valid miRBase versions:  [1] 22.1 22.0 21.0 20.0 19.0 18.0 17.0 16.0 15.0 14.0 13.0 12.0 11.0 10.1
[15] 10.0  9.2  9.1  9.0  8.2  8.1  8.0  7.1  6.0
- Number of species: [1] 277
- Number of database entries among all versions: [1] 291613

Use Cases

Check for valid miRNA name

Valid input values

We have given a set of mature miRNA names miRNAs and would like to check if the names are actual miRNA names that are or were listed in any miRBase version. Our miRNAs have the following values:

> miRNAs = c("hsa-miR-422b", "mmu-mir-872", "gra-miR157b", "cel-miR-56*")

Our first step is to create an instance of the MiRNANameConverter class by calling the constructor and saving it to the variable nc.

> nc = MiRNANameConverter()

Now we check if the names are valid names listed in any of the provided miRBase versions:

> checkMiRNAName(nc, miRNAs)
[1] "hsa-miR-422b" "mmu-mir-872"  "gra-miR157b"  "cel-miR-56*"

Mixed input values

The following set of miRNA names contains valid as well as invalid names.

# "RNU-6" and "bpcv1-miR-B23" are not valid
> miRNAs = c("hsa-miR-422b", "RNU-6", "mmu-miR-872", "gra-miR157b", "bpcv1-miR-B23", "bpcv1-miR-B1")
> nc = MiRNANameConverter()    # Create MiRNANameConverter object
> checkMiRNAName(nc, miRNAs)   # check names
The following miRNAs are not listed in the miRBase repository:
RNU-6; bpcv1-miR-B23
[1] "hsa-miR-422b" "mmu-miR-872"  "gra-miR157b"  "bpcv1-miR-B1"

This time the function prints information for the features that did not pass the check and therefore are not included in the return value.

Translate miRNA names and their sequence

Translating miRNA names to different versions is the most required function. Let us assume we found a paper that shows that the miRNA "hsa-miR-422b" is significantly differentially expressed under certain conditions. Assume further, that we did a similar experiment but this miRNA does not show up in our analysis. Instead we got "hsa-miR-378a-3p" from a recent miRBase release. We see, that the previous paper was released a couple of years ago so there might be a chance that their miRNA could now run under a different name. We apply the translateMiRNAName function with hsa-miR-422b as miRNA parameter and no version. With no given version the function returns the current miRBase version (22.1) by default.

> miRNA.paper = "hsa-miR-422b";
> nc = MiRNANameConverter()             # Create MiRNANameConverter object
> translateMiRNAName(nc, miRNA.paper)   # Translate miRNA names
                   mimat        input           v22.1
MIMAT0000732 MIMAT0000732 hsa-miR-422b hsa-miR-378a-3p

The result shows that these two miRNAs are actually the same. This is because "hsa-miR-422b" is last listed in miRBase version 9.2. In version 10.0 it was named "hsa-miR-378" which in the current version 22.1 runs under the name "hsa-miR-378a-3p".

Another example with more diverse input names is shown below, with the respective console output and the entry in the attribute description.

> miRNAs= c("hsa-miR-128b", "hsa-miR-213", "mmu-miR-302b*", 
            "mmu-miR-872", "ebv-miR-BART5", "bpcv1-miR-B23")
> nc = MiRNANameConverter()              # Create MiRNANameConverter object
> result = translateMiRNAName(           # Translate names
                              nc
                              ,miRNAs
                              ,sequenceFormat = 1
                              ,versions = c(8, 8.1, 18, 21)
                              )        
Following miRNA will be neglected (not listed in miRBase):
bpcv1-miR-B23
Following miRNA is not listed in the current miRBase version 22.1.
hsa-miR-128b

The console output shows us that one of the input values is not a miRNA ("bpcv1-miR-B23") and another is not listed in the current miRBase version ("hsa-miR-128b").

result
                    mimat         input          v8.0          v8.1
MIMAT0000270 MIMAT0000270   hsa-miR-213   hsa-miR-213 hsa-miR-181a*
MIMAT0003373 MIMAT0003373 mmu-miR-302b* mmu-miR-302b* mmu-miR-302b*
MIMAT0003413 MIMAT0003413 ebv-miR-BART5          <NA> ebv-miR-BART5
MIMAT0004934 MIMAT0004934   mmu-miR-872          <NA>          <NA>
                       v18.0            v21.0
MIMAT0000270 hsa-miR-181a-3p  hsa-miR-181a-3p
MIMAT0003373 mmu-miR-302b-5p  mmu-miR-302b-5p
MIMAT0003413   ebv-miR-BART5 ebv-miR-BART5-5p
MIMAT0004934  mmu-miR-872-5p   mmu-miR-872-5p

Attribute `description`

The information, whether a miRNA is OK or not, is stored in form of an attribute of the return value of the function. This data.frame object provides information about every single input value and can be accessed via the attr command followed by 'description'.

attr(result, 'description')
                input.miRNA
bpcv1-miR-B23 bpcv1-miR-B23
ebv-miR-BART5 ebv-miR-BART5
hsa-miR-128b   hsa-miR-128b
hsa-miR-213     hsa-miR-213
mmu-miR-302b* mmu-miR-302b*
mmu-miR-872     mmu-miR-872
                                                               information
bpcv1-miR-B23               This name is not listed in any miRBase version.
ebv-miR-BART5                                                            OK
hsa-miR-128b  This miRNA is not listed in the current miRBase version 22.1.
hsa-miR-213                                                              OK
mmu-miR-302b*                                                            OK
mmu-miR-872                                                              OK

Attribute `sequences`

Sometimes it can be useful to confirm the sequence of miRNAs for selected miRBase versions. In such a case all we have to do is checking out the sequence attribute of our translation result.

attr(result, 'sequence')

##                     mimat         input                    v8.0
## MIMAT0000270 MIMAT0000270   hsa-miR-213  ACCAUCGACCGUUGAUUGUACC
## MIMAT0003373 MIMAT0003373 mmu-miR-302b* ACUUUAACAUGGGAAUGCUUUCU
## MIMAT0003413 MIMAT0003413 ebv-miR-BART5                    <NA>
## MIMAT0004934 MIMAT0004934   mmu-miR-872                    <NA>
##                                  v8.1                    v18.0
## MIMAT0000270   ACCAUCGACCGUUGAUUGUACC   ACCAUCGACCGUUGAUUGUACC
## MIMAT0003373  ACUUUAACAUGGGAAUGCUUUCU  ACUUUAACAUGGGAAUGCUUUCU
## MIMAT0003413 CAAGGUGAAUAUAGCUGCCCAUCG CAAGGUGAAUAUAGCUGCCCAUCG
## MIMAT0004934                     <NA>    AAGGUUACUUGUUAGUUCAGG
##                                 v21.0
## MIMAT0000270   ACCAUCGACCGUUGAUUGUACC
## MIMAT0003373  ACUUUAACAUGGGAAUGCUUUCU
## MIMAT0003413 CAAGGUGAAUAUAGCUGCCCAUCG
## MIMAT0004934    AAGGUUACUUGUUAGUUCAGG

Due to the fact that we called the translateMiRNAName function with the parameter sequenceFormat = 1 the attribute only contains the sequence for each version respectively. If we would like to have miRNA name and sequence combined in one table we can call the translateMiRNAName function with sequenceFormat = 2.

result = translateMiRNAName(            # Translate names
                            nc
                            ,miRNAs
                            ,sequenceFormat = 2
                            ,versions = c(17, 19, 21)
                            )

##     Following miRNA will be neglected (not listed in miRBase):
## bpcv1-miR-B23
##     Following miRNA is not listed in the current miRBase version 22.1.
## hsa-miR-128b

attr(result, 'sequence')

                    mimat         input   v17.0-miRNA
MIMAT0000270 MIMAT0000270   hsa-miR-213 hsa-miR-181a*
MIMAT0003373 MIMAT0003373 mmu-miR-302b* mmu-miR-302b*
MIMAT0003413 MIMAT0003413 ebv-miR-BART5 ebv-miR-BART5
MIMAT0004934 MIMAT0004934   mmu-miR-872   mmu-miR-872
                       v17.0-Sequence      v19.0-miRNA
MIMAT0000270   ACCAUCGACCGUUGAUUGUACC  hsa-miR-181a-3p
MIMAT0003373  ACUUUAACAUGGGAAUGCUUUCU  mmu-miR-302b-5p
MIMAT0003413 CAAGGUGAAUAUAGCUGCCCAUCG ebv-miR-BART5-5p
MIMAT0004934    AAGGUUACUUGUUAGUUCAGG   mmu-miR-872-5p
                       v19.0-Sequence      v21.0-miRNA
MIMAT0000270   ACCAUCGACCGUUGAUUGUACC  hsa-miR-181a-3p
MIMAT0003373  ACUUUAACAUGGGAAUGCUUUCU  mmu-miR-302b-5p
MIMAT0003413 CAAGGUGAAUAUAGCUGCCCAUCG ebv-miR-BART5-5p
MIMAT0004934    AAGGUUACUUGUUAGUUCAGG   mmu-miR-872-5p
                       v21.0-Sequence
MIMAT0000270   ACCAUCGACCGUUGAUUGUACC
MIMAT0003373  ACUUUAACAUGGGAAUGCUUUCU
MIMAT0003413 CAAGGUGAAUAUAGCUGCCCAUCG
MIMAT0004934    AAGGUUACUUGUUAGUUCAGG

Retrieve all miRNAs from different miRBase versions

In cases where one has a dataframe with expression values and miRNA names from a particular miRBase version it can be easier to just ‘join’ a second dataframe with valid miRNA names, such as from v17 to v22.

mirs = data.frame(
        valuex = c(22.3, 21.5, 89.6), 
        mirna = c("hsa-miR-29a", "hsa-let-7f", "hsa-miR-181a*"))
nc = MiRNANameConverter();             # Create MiRNANameConverter object
x = getMirnasForMirbaseVersion(nc, version = 17)
merge(mirs, x, by.x = "mirna", by.y = "v17")

In the case where miRNAs are assigned to multiple accessions, records will be neglected.

Assess most likely miRBase version

Sometimes it is useful to know the miRBase version that a given set of miRNA names is from. In this case one can use the assessVersion function to receive the most likely miRBase version. The following example makes use of the provided example-miRNAs-dataset.

nc = MiRNANameConverter()             # Create MiRNANameConverter object
assessVersion(nc, example.miRNAs)     # Assess most likely miRBase version

##     Input: 355 unique miRNA names
##     The following miRNAs are not listed in the miRBase repository:
## RNU48; RNU44; RNU6B; hsa-miR-517; 4343438
##     miRNAs that could not be found in identified max version 9.0:
## hsa-miR-213; hsa-miR-425; hsa-miR-493
## -> 347 out of 350 valid miRNAs assigned to max version 9.0.
##    version frequency
## 1      9.0       347
## 2      8.2       347
## 3      8.1       347
## 4      9.2       346
## 5      9.1       346
## 6     10.0       276
## 7     10.1       272
## 8     14.0       271
## 9     13.0       271
## 10    12.0       271
## 11    11.0       271
## 12    15.0       270
## 13    16.0       267
## 14     8.0       267
## 15    17.0       266
## 16     7.1       264
## 17     6.0       180
## 18    18.0       133
## 19    19.0       127
## 20    20.0        93
## 21    21.0        92
## 22    22.0        86

The console output shows that from the 384 input names there were 355 unique values. Five of these values are not listed in any miRBase version and were neglected. The return value is a data.frame object with the two columns version and frequency. It is ordered by frequency and version. It shows that 347 out of 350 valid input miRNAs could be assigned to miRBase version 9.0. This is the highest score and therefore the most likely miRBase version of the input mature miRNA names.

Other functions

Save translation results

Sometimes it is useful to save a variable to a file. Taking the translation result from above we can do so by running the following command with default settings:

saveResults(nc, result)

Three files will be saved, the miRNA name translation result, the description and sequence. By default the files are tab-separated files without the parameter quote set to FALSE. Other options can be applied and will be passed on to the utils::write.table function.

Additional information

Database information

The data used in the miRNAmeConverter package is obtained from the miRBaseVersions.db annotation package.