Getting Started

Data can be dropped into the miRiadne features’ forms in different way. The simplest is to directly paste into the form the miRNA names lists copied from external data or files, such as figures, tables or supplementary materials of published papers. It’s also possible to drag and drop tabular data files from a location on own computer directly inside the miRiadne’s form windows. Such tabular data must conform to some simple rules:
  • must be in one of the following format: txt (tab delimited), csv (comma separated value), xls (Microsoft excel). For xls files only the first sheet is considered, others are ignored.
  • miRNA lists do not need a header and miRNA names should be separated by carriage return (new line)
  • if the input file contains expression values (such as CT values from RTqPCR or fluorescence values from microarrays) the files should ideally contain only one column (the first column) for miRNA names and one row (the first row) for samples names. More columns or row can be present but only the first column will be considered as the one containing the input miRNA list and the row just immediately before the expression data will be considered as the one containing the samples names/IDs.

Rosetta Stone - Translation

Translate miRNA name lists to other miRBase or platform annotations.

Translation into other platform's annotation

The Translation feature allows to convert miRNA name lists: the following conversions:
  • from a specific platform’s annotation to the one of a different platform.
  • from a specific miRBase version to a different miRBase version (older or more recent)
  • from a specific platform’s annotation to a specific miRBase version
  • from a specific miRBase version to a specific platform’s annotation
Any combination of paired translation among platforms and miRBases is allowed. Correct match between names and mature sequences is checked and enforced in the translation. Lists of translated names can be downloaded as txt files. Stringent match of miRNA names input is enforced by default, but looser searches can be performed selecting the “Relax search” box before the form. To show how to use this feature we take into consideration the data from Figure1 of Rossi et al.: here 242 miRNAs are shown as the human lymphocyte signature, characterizing the whole 17 cell subsets profiled in the work. miRNA names can be copied from the figure itself and pasted in text format into txt or xls files or directly into the application’s forms. For your convenience we downloaded the list in this txt file: From the methods section of the paper we learn that the miRNA profiling was performed with the Applied Biosystems Taqman Low Density Arrays (TLDA) version 2.0. To convert this signature to the annotation of a different detection platform (for instance the Exiqon Human Panel version 4), the following are the steps to perform:
  • download the files rossi_lymphocytes_signature.txt to your desktop or a convenient location on your computer.
  • drag and drop it into the upper window frame of the Rosetta Stone - Translate page. You can alternatively open the file, select the miRNA list, copy it and paste it into the window frame.
  • Select the “Platform -> Platform” button.
  • As “Origin platform” select the platforms used to generate the data. According to the paper’s methods select “AppliedTLDA:2”.
  • In “Destination platform” select the platforms you want to convert the annotation into: for ths example select “ExiqonHumanPanels:4”.
  • Click on “Convert”
  • Review the result table generated in the page

Result table:

An annotated table appears below listing the miRNAs that have been converted from the annotation used to generate the original signature to the newly selected one. If any searched miRNA is not found in the destination (either platform or miRBase version) a message is displayed in a pink box above the table, and miRNAs not found are highlighetd and listed at the bottom of the table). In this specific example only 227 miRNAs out of the 242 in the original input list are present in the result table: this means that these subset is the one that is also annotated in the destination platform. The result table displays the names correspondence between the two platforms and the mature sequences, and can be downloaded as a csv file; the unmatched miRNAs (shaded in pink at the bottom of the table) are not included into the downloadable file.

Translation into other miRBase version

Suppose you want now to see if all 242 miRNAs in the original signature can be converted into a more recent miRBase annotation. The original files were generated with Applied Biosystems TLDA version 2.0, which is annotated after miRBase version 10. (see the Platform page for details). To convert this signature into the annotation of a more recent miRBase, for example the miRBase version 19, the following are the steps to perform:
  • download the files rossi_lymphocytes_signature.txt to your desktop or a convenient location on your computer (if you didn’t do it already).
  • drag and drop it into the upper window frame of the Rosetta Stone - Translation page. You can alternatively open the file, select the miRNA list, copy it and paste it into the window frame.
  • Select the “Platform -> miRBase” button.
  • As “Origin platform” select the platforms used to generate the data. According to the paper’s methods select “AppliedTLDA:2”.
  • In “Destination miRBase” select the version you want to convert into: for this example select “19.0”.
  • Click on “Convert”
  • Review the result table generated in the page

Result table

An annotated table appears below listing the miRNAs that have been converted from the annotation used to generate the original signature to annotation of miRBase version 19. In this specific example the result table shows that all 242 input miRNAs can be converted in 237 miRNA annotations according to version 19 of miRBase. Looking at the Destination column (miRBase v19.0) we see that miRNAs such as hsa-miR-516a-3p can be converted in more than one annotation (hsa-miR-516b-3p\hsa-miR-516a-3p) and other miRNAs such as hsa-miR-139-5p display an empty cell, meaning that they were dropped out in (or before) miRBase v.19. Such cases are shaded in color for easier spotting in the table. With this search it’s possible to see that, among the 242 miRNAs in the original list, when updated to miRBase 19, 2 miRNAs have are annotated with multiple names, while 5 miRNAs were dropped out or were subjected to some sequence-related changes (hsa-miR-139-5p, hsa-miR-517b, hsa-miR-801, hsa-miR-886-3p, hsa-miR-886-5p). To have a complete picture on what happened to these 5 dropped miRNAs, and at what stage of nomenclature evolution they were retired from miRBase, you can use the Time Warp function on those miRNAs.

Rosetta Stone - Overlap

Find miRNAs that are commonly detected by different platforms or are conserved in different miRBase versions.

This function allows to find the miRNAs in common between two different detection platforms (i.e. present on both platforms with the same mature sequence regardless names variations), between two miRBase versions or between platforms and miRBase versions. Any difference in names for same-sequence miRNAs in the overlap is highlighted. Suppose you want to know how many and which miRNAs are detected by both the Exiqon Human Panel version 3 and Affymetrix miRNA Array version 2. If you would be tempted to simply find the overlap by miRNA names annotation with a text-based methods you would realize that, since the two platforms refers to different miRBases, the common miRNA names to both platforms are 365. Nonetheless, the miRNA in common are much more than 365: how many and which ones is easily found using the Rosetta Stone - Overlap function in this way:
  • click on the “Rosetta Stone - Overlap” page
  • choose the “Platform ^ Platform” button
  • select the first platform (AffyMirArray2)
  • select the second platform (ExiqonHumanPanels:3)
  • click on “Convert”
  • Review the result table
As you can see the miRNAs in the result table are much more than 365 (they are exactly 748), and the table reports the correspondence between names in the two platforms and the relative mature sequence.

Rosetta Data - Update

Update your own miRnome datasets with more recent annotation.

This function allows to update a miRNA dataset to a more recent version of the miRBase (or to the annotation of a different platform), keeping the original expression values from the dataset. As an example, suppose you want to update the lymphocyte miRnome dataset from Rossi et al. (2011) to re-analyze it with newer nomenclature, thus excluding any dropped out miRNA or taking into account any name change. This function allows to update the miRNA names keeping the corresponding data (expression values) from the original dataset in the downloadable newly annotated table. The data for this example can be retrieved from the original paper:
  • Supplementary Spreadsheet of Rossi et al. (2011) contains all the CT values for the miRnome profiling of 17 human lymphocyte cell subsets. From the “Methods” section of the paper we learn that the profiling was performed with the TaqMan Low Density Arrays v.2.0 from Applied Biosystems. The dataset can be downloaded from the paper’s page, for your convenience you can also download it here rossi_lymphocytes_mirnome.txt.
To do so follow these steps:
  • download the files rossi_lymphocytes_mirnome.txt to your desktop or a convenient location on your computer.
  • drag and drop it into the upper window frame of the Rosetta Stone - Translation page. You can alternatively open the file, select the entire data matrix, copy it and paste it into the window frame.
  • To update the dataset’s annotation to the most recent miRBase, click on “Platform->miRBase”
  • As “Origin platform” select the platforms used to generate the data. According to the paper’s methods select “AppliedTLDA:2”.
  • As “Destination miRBase” select the most recent one, i.e. “21.0”
  • Click on “Convert”
  • Review the generated result table.

Result table:

annotations (miRNA names) of the original dataset are reported side by side with the updated annotations (name and mature sequence) and with the “Data” column containing all the original dataset’s expression value. The table can be downloaded as a csv file, where data values can be displayed in separate columns too, thus allowing further elaboration and/or analyses. The updated annotation can be a blank cell (for dropped out miRNAs, no longer present in the miRBase) or multi-named slash separated annotation (for miRNAs with different names but with identical mature sequence). miRNAs that cannot be found in the destination annotation (either a more recent miRBase version or a different platform) are reported shaded in red at the bottom of the table and a warning is automatically displayed before the table. Please note that typically annotations from control probles, non miRNA-probes or endogenous controls (i.e. any annotation which is by definition not included in the miRBase) cannot be recognized and thus will trigger the above mentioned warning.

Rosetta Data - Intersection

Find the intersection of heterogeneously annotated miRNA datasets.

The “Intersection” feature allows to find common miRNAs from two human profile datasets (data matrices with miRNA expression values/rows in a number of samples/columns) generated by two different profiling platforms. The result is a third dataset, made up by common miRNAs only, with annotations updated to the selected (miRNA names). The newly generated dataset contains all the miRNAs present in both original datasets according to the mature sequence (i.e. that have the same sequence), regardless any change to the miRNA names that could have been occurred. The new data matrix can be then downloaded as a csv file. To show how to use this feature we take into consideration the data in Skogberg et al. (2013) and Vickers et al. (2011). Both articles deal with miRNA content of human exosomes: while Skogberg et al. mainly describes thymic exosomes, Vickers et al. describes high-density lipoproteins (HDL) vesicles and exosomes are profiled for comparison. Here we illustrate how to compare the exosomes’ miRNA profile from Skogberg et al. to the HDL vesicle profile from Vickers et al. The original data that can be retrieved from the original papers. They are:
  • TableS4 of Skogberg et al. shows the miRNAs detected in Human Thymic Hexosomes. We limit the miRNA list to the miRNAs annotated with the “hsa” prefix. From the “Methods” section of the paper we learn that the profiling was performed with the “Agilent human miRNA microarray release 14.0” platform.
  • Supplementary Table 2 from Vickers et al. show the miRNAs detected in human HDL vesicles. From the “Methods” section of the paper we learn that the profiling was performed with the TaqMan Low Density Arrays v.2.0 from Applied Biosystems.
miRNA lists are given as supplementary material in pdf format and they can be copied and pasted in text format into txt or xls files or directly into the application’s forms. For your convenience we downloaded them and here you can see and retrieve the files containing the above mentioned miRNA lists: Please note that miRNA signature lists are given with miRNA names only, and no sequence is provided by the original publications: only a reference to the platform used to generate the data. Steps to compare human exosome miRNAs from Skogberg et al. to human HDL vesicles miRNAs from Vickers et al.: download the files skogberg_exosomes_mirs.txt and vickers_hdlvesicles_mirs.txt to your desktop or a convenient location on your computer. drag the first (skogberg_exosomes_mirs.txt) and drop it into the upper window frame of the Rosetta Data - Intersect page. You can alternatively open the file, select the miRNA list, copy it and paste it into the window frame. drag the second (vickers_hdlvesicles_mirs.txt) and drop it into the lower window frame of the Rosetta Data - Intersect page. You can alternatively open the file, select the miRNA list, copy it and paste it into the window frame. Select the platforms used to generate the data. According to each paper’s methods select “AgilentHSA:14” for the first and “AppliedTLDA:2” for the second. Click on “Submit” Review the result table generated in the page

Result table:

An annotated table appears below the forms containing 15 miRNAs: these miRNAs are those shared by the two original platforms that belongs to both the dataset with consistent sequence correspondence. The result table of the Rosetta Data - Intersection reports the input miRNA lists (first two columns), the name annotation as reported in the most recent miRBase (third column) and the relative mature sequence (fourth column). This output allow to see if any miRNA name changed since the original dataset were generated and annotated: in our case three miRNAs changed their names: hsa-let-7b, hsa-miR-23a*, hsa-miR-425* and hsa-miR-766. If the original miRNA lists, the same that has been dropped in the Rosetta above, had been compared by a simple name-based overlap, the common miRNAs would have been 16 and not 15 since the additional hsa-miR-923 would have been picked up too. This miRNA is a drop-out and Rosetta checks for cases such these and automatically purge them. If you want to know how many miRNAs among those detected in HDL vesicles by Vickers et al. are also detectable by another detection platform, for example the Exiqon Array 7th generation, you can use the function Rosetta Stone - Translation linked to TimeWarp Tutorial with the vickers_hdlvesicles_mirs.txt. You can also ask how many miRNAs had their names changed or simply dropped out among those of the original dataset. To answer this question you can use the TimeWarp Tutorial function on the very same datasets.

Time Warp

Browse the miRNA annotations’ evolution.

This function allows to track miRNA names evolution (name changes, sequence annotation changes, survival and/or death) through the miRBase versions from number 10 to the most current one. Name changes, new miRNA entries and old dropped out miRNAs are highlighted. Mature sequence for each entry is also appended. We learn from the tutorial for Rosetta Data - Intersection that 5 miRNAs, belonging to the 242 miRNAs lymphocyte signature obtained by Rossi et al. (2011) (see Figure1a of the paper) were subjected to some kind of changes, since they were generated with a miRBase v.10 compliant platform and couldn’t be re-annotated with a more recent version of miRBase. To find out what happened to these miRNAs click on TimeWarp page, digit their names (hsa-miR-139-5p, hsa-miR-517b, hsa-miR-801, hsa-miR-886-3p, hsa-miR-886-5p) or drag and drop the file rossi_lymphocytes_changed.txt containing their names into the TimeWarp form; leave the default parameters and click search, then review the result table generated below.

Result table:

the five miRNAs are displayed with their MI-id (precursor ID) and MIMAT-id (mature form ID) from miRBase. The searched names are reported in the column corresponding to the miRBase versions: in this case we didn’t specify a miRBase version (in the “Select miRBase Version” option) and search was performed starting from the older version considered in the application (version 10). Each column correspond to a miRBase version, and in each row is shown the annotation’s evolution for a single miRNA. A green dot means that nothing changed, while a red dot means that the miRNA was dropped out from miRBase. If a name change occurs, the new name is displayed in correspondence of the miRBase version in which the change was introduced. The last three columns are for the mature sequence, the arm in the precursor hairpin and the length of mature form. Changes in the 3’ or 5’ nucleotides (addition or lack of single nucleotides) are highlighted in the mature sequence column. In this specific example we can observe the following changes:
  • hsa-miR-139-5p is substantially unchanged in name annotation, but the sequence changed undergoing a 3’ offset, and [-”U”] notation means that older sequence lacks a “U” that is present in the current miRBase annotation.
  • hsa-miR-517b changed its name to hsa-miR-517b-3p in miRBase v.18; additionally older sequence lacks an “A” at the 5’ end and has an additional “U” at the 3’ end, compared to the sequence in the current miRBase annotation.
  • hsa-miR-801 was dropped out in miRBase v.11
  • hsa-miR-886-3p and hsa-miR-886-5p were dropped out in miRBase v.16
Much longer miRNA lists can be searched with the TimeWarp function, for instance entire miRNA signatures, to visually inspect any change that might have occurred during time, i.e. across miRBase versioning. The default parameters perform searches starting from miRBase version 10 and for human miRNAs (with the “hsa-” prefix), but more recent version of the miRBase can be selected to start the search with, and searches for other specie can be performed too. Relaxed searches can be done checking the “Relax search?” option. With a relaxed search partial names can be used as input, such as “hsa-let” or “hsa-miR-3” (resulting in all miRNAs with a “let” notation or whose first number is a “3”, respectively). Combining the relaxed search with the “blanck” option in the “Select specie” choice, all miRNAs with a specific name can be searched across the miRBase and species.

References

These are the references cited in the tutorial page

  • Rossi et al (2011) Distinct microRNA signatures in human lymphocyte subsets and enforcement of the naive state in CD4+ T cells by the microRNA miR-125b. Nat. Immunol. 12(8):796-803.
  • Skogberg et al. (2013) Characterization of Human Thymic Exosomes. PLoS One 8(7):e67554.
  • Vickers et al. (2011) MicroRNAs are transported in plasma and delivered to recipient cells by high-density lipoproteins. Nat. Cell Biol.13(4): 423-433.