Skip to content

miRBase 17 released

miRBase 17 is finally released. The README has stats on the numbers of new sequences (of which there are again many). We’ve added a bunch more deep sequencing data, and there’s more on the way. There are also two important changes in nomenclature, about which we’ll write more shortly. From the README:

The -as nomenclature (previously used to designate a miRNA that is
antisense to another) is discontinued, and a small number of sequences
are renamed for this reason.

The first steps to retire the miR* designation are also taken here.
Mature sequence from all Drosophila melanogaster precursors are now
designated -5p and -3p. Many minor mature products have been added,
and others are renamed.

Some people have been asking if the views of deep sequencing data are available in bulk. We’re working on that, and we’ll post here when you can get them.

Comments, questions, abuse, praise all welcome.

Posted in data update, releases.

12 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Nicola Soranzo says

    Hi Sam,
    have you already submitted the new microRNAs to NCBI Gene and/or HUGO/other species nomenclature organizations? Do you have a timeline on when the new miRNAs will be crosslinked with the relative entries?
    Thanks in advance for the info, and happy new release!

  2. sam says

    Actually we don’t usually submit data to those places, rather they grab the data from us on their timescales. MGI usually does this very quickly, and HUGO and Refseq also pick up changes and feedback conflicts etc. Once those resources point to specific miRBase entries, we can then grab those pointers to reciprocate the links. That means links from miRBase get updated in the following release cycle, so there is always a bit of a lag.

  3. Wen-Ching Chan says

    Hi Sam,
    I found that three miRNAs belong to two families in miRBase 17. (not sure whether the same situation happened in previous releases.)
    They are
    #auto_mirna mirna_acc mirna_id auto_species name auto_prefam prefam_acc prefam_id
    70887 MI0007261 hsa-mir-103b-1 22 Homo sapiens 5116 MIPF0000024 mir-103
    70887 MI0007261 hsa-mir-103b-1 22 Homo sapiens 5717 MIPF0000644 mir-103-as
    70888 MI0007262 hsa-mir-103b-2 22 Homo sapiens 5116 MIPF0000024 mir-103
    70888 MI0007262 hsa-mir-103b-2 22 Homo sapiens 5717 MIPF0000644 mir-103-as
    75138 MI0011626 dps-mir-iab-8 47 Drosophila pseudoobscura 5242 MIPF0000151 mir-iab-4
    75138 MI0011626 dps-mir-iab-8 47 Drosophila pseudoobscura 5918 MIPF0000854 mir-iab-8
    Do you have any documentation related to these assignments of multiple families for a miRNA?
    Many Thank.

    • sam says

      Thanks! That was a bug with the database dump. I’ve fixed that now on the FTP site — you’ll want to re-download the mirna_prefam and mirna_2_prefam tables. The miFam.dat family definition file was correct and unaffected.

  4. Wen-Ching Chan says

    Hi Sam,
    bdi-MIR171d(MI0011569, MI0018090), bdi-MIR169a(MI0011564, MI0018120), bdi-MIR408(MI0011560, MI0017941) are three ‘dead’ miRNAs but without removing from mature.fa, hairpin.fa, and miRNA.str.
    Does that the reason why 16,772 miRNAs are reported in the change log but only 16,769 miRNAs could be found in database_files/mirna.txt?
    Many thanks.

  5. Jennifer Williams says

    Release 17 is nice & the “What’s in a name?” post really helped me to understand the updated names & why you changed the format.

    Could the name changes have broken some of the target data links? I am asking because and MIMAT0000398 both have links to target data at various databases, but several of the links return “no hits/data” messages.

    Thanks for any help/info!

    • sam says

      Yep, you’re right. We’re currently building some sets of those target links in a dumb way. The large-scale name changes have highlighted that. We’ll fix them as soon as we can. I guess the stop-gap solution is to use the search interface in the target resource itself, stripping off the -5p or -3p part of the ID. For example, searching for dme-mir-307 in Targetscan works.

  6. Fengshou Ma says

    Hi Sam,
    Where is the miRBase_v17.fa file?
    I can’t find it.

  7. sam says

    The fasta format sequence download are split into hairpin.fa and mature.fa on the FTP site.

  8. Bino John says

    Hello Sam,
    I emailed you earlier directly, but didn’t get a reply.I probably used the wrong email from web. Could you please clarify what is the current definition of a microRNA that mirBAse uses? Does mirbase accept any abundant small rna that has a stemploop preccursor as a microRNA?

  9. sam says

    Hi Bino

    Sorry for the delay — desperately trying to get a release out! The criteria for sensible annotation of novel miRNAs (particularly by deep sequencing) are evolving fast. We wrote a little about this in the the last NAR update paper (Kozomara and Griffiths-Jones, NAR 2011 39:D152-D157). It isn’t quite as simple as any short sequence in a hairpin. The patterns of reads across a locus can provide good evidence for a miRNA annotation. We’re not far away from requiring that both miR and miR* sequences are supported by significant numbers of reads, with consistent 5′ ends and 2nt overhang. We’re not exactly imposing this yet, but I think that’s the only sensible way to go now. We are however going to start flagging entries that do and don’t have this level of evidence.