ZENBU Howto

From Wiki
Jump to navigationJump to search

ZENBU user documentation

The ZENBU system is under development by Jessica Severin from the OSC. Zenbu is documented on http://fantom.gsc.riken.jp/zenbu/wiki/, and in addition we use these FAQs to guide you. Please ask your questions here.

  • FAQs
    • How can I find my genes of interest?
    • How can I see an overview of the data available in the ZENBU system?
    • What is the difference between "Tracks", Annotation data and Expression experiments?
    • How can I see the expression of the region of my interest?
    • What is the difference between the green and the red bars?
    • Can I re-arrange the tracks?
    • Can I temporarily hide (e.g. collapse) tracks?
    • What is the unit for expression on the y-axis of the tracks?
    • Can I change the expression unit?
    • Can I change the hight of one track?
    • Can I change the withs of the tracks?
    • Can I move the bottom window with expression bars?
    • What is a configuration / config? Please explain the concept (incl. how to save configs and make them visible to other users).
    • Sometimes the response of the website takes time. Is there an indicator showing if the request is still in progress and I should wait?
    • Can I upload my own data to ZENBU? Will it be private or share with all other users?
    • What is my advantage to register as individual user in ZENBU over the default guest account?

Some questions about ZENBU and FANTOM5 data sharing from Leonard Lipovich

Dear Leonard,

Thank you so much for your positive comments and good questions. This will be a long email since your questions are a bit complicated and require a bit of detail to answer.

What is the easiest way to display hg19 shared annotation tracks (e.g. lncRNAome) in Zenbu? And is this available only under the collaborator / data sharing login (not the general F5 login)?

Currently I can see the lncRNAome annotation data without login (guest user). I don't know the full history of this particular dataset but it was uploaded by some user and shared. The general procedure for sharing data with all FANTOM5 members is to

  1. login to ZENBU
  2. go to the user/collaborations section and make sure you are a member of the collaboration called "FANTOM5_general_sharing". If not please click the "request to join" button. I manage this collaboration so I will need to approve each user's request. Unfortunately the system is not advanced enough yet to email me when I need to check, so please feel free to send me personal email and I can more quickly approve your request.
  3. upload your data files
  4. go to the user/my data section and for each data file click the [share] button. A panel should appear with at least the "FANTOM5_general_sharing" collaboration shown. Click the [share] button next to the the "FANTOM5_general_sharing" collaboration.

At this point one's data file is now made "public" to all FANTOM5 guest users. Since the entire FANTOM5 website is protected under the FANTOM5 website password, it is OK to make any data available to FANTOM5 guest users. But of course if anyone wants to create private collaborations, they can, but then users must login to the ZENBU system to see the data. Only data shared via "FANTOM5_general_sharing" is made available to "guest" (no login) users.


Is the 409-sample mouse hCAGE CTSS update (June 1, 2011) ready for use? We are studying gene structure nonconservation between human and mouse; viewing evidence of bidirectional transcription (or lack thereof) at specific orthologous promoters would be useful. My concern is that - unless I missed something - there was no update on the F5 ML or on the Wiki about this very significant new mouse release. Wiki overview still refers to just the 30 mouse primary samples.

It appears ready to me, but I should refer this question to Marina and Kawaji-san. They are managing the official FANTOM5 data releases and official data loading into the FANTOM5 ZENBU. This might be a simple mistake of not updating the wiki page with the new link. It also appears to be 407 experiments not 409 for some reason.

Here is Marina's all-experiments-1-track 407-sample mouse hCAGE CTSS update (June 1, 2011) (6 tracks in view) 

 https://fantom5-collaboration.gsc.riken.jp/zenbu/gLyphs/#config=vzLEJkOeMZ7lJW4eDwC3qB;loc=mm9::chr6:85458240..85466410

Here is Marina's 1-experiment-1-track 407-sample mouse hCAGE CTSS update (June 1, 2011) (>400 tracks in the view) 

 https://fantom5-collaboration.gsc.riken.jp/zenbu/gLyphs/#config=beTZAt9OKNTk6JnyjRFNJC

personally I find the 400 track style views very difficult to work with (too much scrolling, very slow loading). I prefer to use a single expression track pooled with all experiments and use the interactive region selection option to explore spatial variation within the track. If I understood the biological questions better I might create different expression tracks with different pooling "concepts" like maybe one track being all cell-lines and another track being all primary cells so I could easily see if there are differences in a nice compact two track manner. Since ZENBU allows users to create new visualization tracks using the already loaded data by using different pooling filters, this is easy without having to preprocess and reload data set pools.


When clicking on a tag cluster in Zenbu, is it possible (or will it be possible in the future) to display the actual sequences of the tags which comprise that cluster? As an annotator, I think that the ability to go back to an original tag (and BLAST or BLAT it vs. the genome or transcriptome if desired) may provide a valuable additional means of QC. This functionality (show actual tag sequence) was available in the F5 CTSS viewer if I recall correctly.


The ZENBU system already provides some of this functionality, but it requires that particular data to be loaded. Currently the tag clusters have been loaded without the original sequence tags (smaller data files, faster loading and query times). If this is a feature that many people would be interested in, then we would need to

  • reprocess the tag clusters to retain the original sequence tags
  • reload the tag clusters with this additional data present
  • possibly modify the current ZENBU interface to provide a better user experience accessing this data. Might be 1 line of code to allow that particular "metadata" tag to be displayed in the existing hover-info-box.


Would it be possible to institute a master track control in Zenbu (similar to that in UCSC) so that we can click experiments or experiment groups (e.g. sets of related tissues from the human 676-sample data) On/Off before displaying the tracks? The current functionality is to display all the tracks first, which takes a very long time ("track reloading"... message) and therefore hinders, in our experience, the use of this otherwise very sophisticated browser.


ZENBU differs from UCSC in a very fundamental concept. With UCSC the data content is curated by UCSC (and thus the interface can be updated and tweaked based on the limited curated content) With ZENBU, data content is managed/curated by all the users of the system, thus the interface must be flexible to deal with both the dynamic nature of the content and the large number of possible data sets and visualization tracks possible.

But ZENBU does provide the concept of a "saved visualization track". This is similar to a "saved visualization view" (all tracks) but is just for a single track. Any user can save a track, and anyone can add a saved track into an existing view.

For example I just saved the "Helicos ctss 407 sample pool" visualization track into the system so anyone can access this preconfigure and add it into any mouse mm9 view. You can see this preconfigured track here (search mm9 helicos) 

  https://fantom5-collaboration.gsc.riken.jp/zenbu/dex/#section=Tracks

An example of adding this track would be to follow something like this.

  1. open up a view you like (example basic mm9 view)
  https://fantom5-collaboration.gsc.riken.jp/zenbu/gLyphs/#config=FMWyUSME4BGkTZIkQB9Gew;loc=mm9::chr7:52251732..52259514
  1. in the main search box at the top search for "mm9 helicos ctss". This is where people usually type in gene names to find gene locations, but it is multi-functional.
  2. the second search section is labeled "track configs" and one can see the "Helicos mm9 ctss (407 sample pool)" track. One could also just search with "mm9" and see all predefined mouse tracks (10 currently as I write this). One can hover of the track-configs (just like genes or views) and see additional description about the track, and maybe refine the query.
  3. click the track you want to add, and this track will be loaded and added to the bottom of your view.

This is not the same as the UCSC master track control (many sections, many buttons and many pull downs), but it provides a similar nature of functionality in a simple search style interface. To remove tracks from one's view use the "red close X" control on right side of each track. Or hide the track with the "blue circle/line" control on the left side of each track.

Of course I am always thinking of new interface ideas and ways to make the existing interface more ergonomically useful.

Retrieved from "http://fantom5-collaboration.gsc.riken.jp/wiki/index.php?title=ZENBU_Howto&oldid=2665"