[maker-devel] maker2jbrowse implementation

Robert Buels rbuels at gmail.com
Thu Jul 11 12:27:10 MDT 2013


These days, it's better to use direct BAM access.  bam-to-json.pl still 
works, but the direct BAM access is much better.

See 
http://gmod.org/wiki/JBrowse_Configuration_Guide#Alignment_Tracks_.28BAM.29


Robert Buels
Lead Developer
JBrowse - http://jbrowse.org

On 07/11/2013 02:23 PM, Carson Holt wrote:
> If you have a previous Jbrowse with already populated tracks you can
> just, run maker2jbrowse in that directory instead of the new Jbrowse
> directory, then the maker features will be added there.  Make a backup
> copy of the directory before you do this of course incase you need to
> roll back the changes.
>
> Alternatively you will have to add those to the new Jbrowse directory
> using bin/bam-to-json.pl and following the instructions here -->
> http://gmod.org/wiki/JBrowse_Tutorial_2012#BAM_data
>
> I guess Robert can chime in here as well, since we have now moved
> outside of MAKER scripts and fully into JBrowse.
>
> --Carson
>
>
> From: V Wong <vwong at umn.edu <mailto:vwong at umn.edu>>
> Reply-To: <vwong at umn.edu <mailto:vwong at umn.edu>>
> Date: Thursday, July 11, 2013 1:58 PM
> To: Carson Holt <carsonhh at gmail.com <mailto:carsonhh at gmail.com>>
> Cc: Ignazio Carbone <icarbon at ncsu.edu <mailto:icarbon at ncsu.edu>>,
> <maker-devel at yandell-lab.org <mailto:maker-devel at yandell-lab.org>>,
> Robert Buels <rbuels at gmail.com <mailto:rbuels at gmail.com>>
> Subject: Re: [maker-devel] maker2jbrowse implementation
>
> Hi Carson,
>
> Ah, great!  After moving the contents of ../JBrowse/data/ to our browser
> directory, the Maker tracks are now on view.
>
> However, I'm having some trouble getting our older tracks (bam
> alignments and coverage tracks) into the same browser.  Neither 1.
> copying and pasting the tracks from the previous trackList.json to the
> new one nor 2. running add-track-json.pl <http://add-track-json.pl> is
> doing the job.  Is it possible to merge these tracks into one browser?
>
> Best,
> Val
>
>
> On Thu, Jul 11, 2013 at 6:44 AM, Carson Holt <carsonhh at gmail.com
> <mailto:carsonhh at gmail.com>> wrote:
>
>     You next put the entire JBrowse directory into the /var/www/html/
>     directory or wherever you have apache set up to load from (if it's
>     not already there).  Then it will be accessible as a webpage
>     (browser is inside the JBrowse directory).
>
>     Example /var/www/html/Jbrowse/
>     Webpage --> http://localhost/Jbrowse/
>
>     --Carson
>
>     From: V Wong <vwong at umn.edu <mailto:vwong at umn.edu>>
>     Reply-To: <vwong at umn.edu <mailto:vwong at umn.edu>>
>     Date: Thursday, July 11, 2013 1:10 AM
>     To: Carson Holt <carsonhh at gmail.com <mailto:carsonhh at gmail.com>>,
>     Ignazio Carbone <icarbon at ncsu.edu <mailto:icarbon at ncsu.edu>>
>     Cc: <maker-devel at yandell-lab.org
>     <mailto:maker-devel at yandell-lab.org>>, Robert Buels
>     <rbuels at gmail.com <mailto:rbuels at gmail.com>>
>
>     Subject: Re: [maker-devel] maker2jbrowse implementation
>
>     Hi Carson,
>
>     Thanks!  I followed your instructions, and maker2jbrowse, and it
>     successfully populated the ../JBrowse/data/tracks directory with
>     RepeatMasker, Augustus, and GeneMark predictions.
>
>     I'm not entirely sure how to get these numerous tracks into the
>     browser, but that may be something to work out with JBrowse rather
>     than Maker.
>
>     Best,
>     Val
>
>
>     On Tue, Jul 9, 2013 at 10:57 AM, Carson Holt <carsonhh at gmail.com
>     <mailto:carsonhh at gmail.com>> wrote:
>
>         Hi Val,
>
>         I've attached new files for you to test.  These will be bundled
>         in the
>         next maker release pending successful tests (~ 1 week).
>
>         Instructions:
>         1. Place maker2jbrowse in your .../maker/bin/ directory and
>         .../maker/src/bin/ directory (this is just to ensure you don't
>         accidentally overwrite it if you reinstall at any time).
>         2. Place maker.css in the .../maker/GMOD/Jbrowse/ directory.
>         3. You must be in the jbrowse installation directory when running
>         maker2jbrowse.  Run the script with the -d flag and the
>         datastore index
>         file.
>
>         This new version will assign colors by copying maker.css to the
>         jbrowse
>         installation as apposed to replacing the genome.css file, and it
>         will also
>         edit the genome.css file to add the import call for maker.css.  In
>         addition I have updated command line flags in the wrapper to do
>         what is
>         currently expected by flatfile-to-json.pl
>         <http://flatfile-to-json.pl>.
>
>         Could you test it out and let me know how it goes.
>
>         Thanks,
>         Carson
>
>
>
>         On 7/3/13 3:26 PM, "Carson Holt" <carsonhh at gmail.com
>         <mailto:carsonhh at gmail.com>> wrote:
>
>         >Like I said that css file is ~1 year old.  Once upon a time the Jbrowse
>         >documentation had you modify the genome.css file to create your own
>         >tracks.  Version 1.2.1 looks like it is the last version where the
>         >genome.css file that maker provided works. I can still get simple test
>         >GFF3 tracks to load with current Jbrowse and the maker2jbrowse script.
>         >But I need to fix some things, because they are all indistinguishable from
>         >each other (so it is not the desired behavior).
>         >
>         >Indepent of that issue though, to regenerate the original error, I do the
>         >following -->
>         >    maker2jbrowse
>         >~/Developer/maker/trunk/data/dpp_contig.maker.output/dpp_contig_master_dat
>         >a
>         >store_index.log
>         >    ERROR: No track information for source
>         >'dpp_contig_datastore/05/1F/contig-dpp-500-500/'
>         >
>         >Basically you are missing the '-d' flag (whiteout that the script is
>         >expecting a list of GFF3 files).
>         >
>         >So this works -->
>         >    maker2jbrowse -d
>         >~/Developer/maker/trunk/data/dpp_contig.maker.output/dpp_contig_master_dat
>         >a
>         >store_index.log
>         >
>         >
>         >Of course all tracks look the same though per the original issue I
>         >mentioned.
>         >
>         >Thanks,
>         >Carson
>         >
>         >
>         >
>         >Some versions of Jbrowse still load fine with the script as is Jbrowse
>         >1.3,
>         >
>         >On 7/3/13 12:39 PM, "Robert Buels" <rbuels at gmail.com <mailto:rbuels at gmail.com>> wrote:
>         >
>         >>Regarding the genome.css file that's in GMOD/JBrowse, I'm not sure what
>         >>you're doing with that, but it appears to be a fork of a very old
>         >>JBrowse main css file.  If that's overriding or being copied over the
>         >>JBrowse released CSS, it will break pretty much everything.  It would be
>         >>like me randomly replacing one of MAKER's Perl modules with my own old
>         >>forked copy.
>         >>
>         >>If what you're trying to accomplish is to add some custom styles, the
>         >>best way to do that would be to do one of:
>         >>
>         >>* add just your custom feature styles on theflatfile-to-json.pl <http://flatfile-to-json.pl> command
>         >>line
>         >>* add a "css": { "url": "http://url/of/your/custom.css" } to the
>         >>   JBrowse trackList.json someplace in the workflow (easy to add
>         >>   things to JSON files with JSON.pm) that loads your custom CSS
>         >>   (containing only your custom feature styles!)
>         >>* add all of the custom styles to the trackList.json as a
>         >>   string: "css": "string { of: lots and lots of; css: properties; }"
>         >>
>         >>
>         >>Robert Buels
>         >>Lead Developer
>         >>JBrowse -http://jbrowse.org
>         >>
>         >>On 07/03/2013 12:02 PM, Carson Holt wrote:
>         >>> I think you may not have it set up right.  If you run the example data,
>         >>> you can then use that.  For me it populates the JSON data tracks
>         >>> information, but nothing displays.  I think the problem is with the
>         >>>maker
>         >>> genome.css.  It is very out of date.  I'll do some work on it and then
>         >>>get
>         >>> your feedback.
>         >>>
>         >>> Thanks,
>         >>> Carson
>         >>>
>         >>>
>         >>>
>         >>> On 7/3/13 11:45 AM, "Robert Buels" <rbuels at gmail.com <mailto:rbuels at gmail.com>> wrote:
>         >>>
>         >>>> After looking at the maker2jbrowse source (thanks for giving me
>         >>>>access),
>         >>>> there seems to be some kind of problem with the way maker2jbrowse is
>         >>>> parsing the MAKER datastore index, leading to the script complaining
>         >>>>of
>         >>>> 'no track information'.
>         >>>>
>         >>>> This would probably be easier for a MAKER developer to look into.
>         >>>>
>         >>>>
>         >>>> Robert Buels
>         >>>> Lead Developer
>         >>>> JBrowse -http://jbrowse.org
>         >>>>
>         >>>> On 07/02/2013 12:27 PM, Carson Holt wrote:
>         >>>>> The Jbrowse codebase tends to undergo very rapid alteration (much of
>         >>>>> which is not backwards compatible with other Jbrowse versions), as a
>         >>>>> result the script maker2jbrowse has had to evolve with Jbrowse (and
>         >>>>>will
>         >>>>> not work with all versions).  I'm not sure if the script is currently
>         >>>>> behind or ahead of Jbrowse version 1.9.3 (I'm guessing behind). I can
>         >>>>> have it working with whatever the most current repository is if you
>         >>>>>were
>         >>>>> to give me a week, or you can try loading the GFF3's using Jbrowse's
>         >>>>>own
>         >>>>> loader (tutorial here -->
>         >>>>>http://gmod.org/wiki/JBrowse_Tutorial_2012).
>         >>>>>    Of course this tutorial is also likely to be somewhat out of date
>         >>>>> :-( But the 2013 version is supposed to be up and running as part of
>         >>>>> this years GMOD summer school in ~3 weeks.  They use MAKER generated
>         >>>>> data for the tutorial, so it should be relevant to what you are
>         >>>>>doing.
>         >>>>>
>         >>>>> Thanks,
>         >>>>> Carson
>         >>>>>
>         >>>>>
>         >>>>>
>         >>>>> From: V Wong <vwong at umn.edu <mailto:vwong at umn.edu> <mailto:vwong at umn.edu
>         <mailto:vwong at umn.edu>>>
>         >>>>> Reply-To: <vwong at umn.edu <mailto:vwong at umn.edu> <mailto:vwong at umn.edu
>         <mailto:vwong at umn.edu>>>
>         >>>>> Date: Tuesday, July 2, 2013 11:38 AM
>         >>>>> To: <maker-devel at yandell-lab.org <mailto:maker-devel at yandell-lab.org>
>         >>>>><mailto:maker-devel at yandell-lab.org <mailto:maker-devel at yandell-lab.org>>>
>         >>>>> Subject: [maker-devel] maker2jbrowse implementation
>         >>>>>
>         >>>>> Hi Maker team,
>         >>>>>
>         >>>>> I'm trying to get Maker output into JBrowse (version 1.9.3) on our
>         >>>>> server, using the -d option for the index.log file.  However, I am
>         >>>>> getting this error for each contig:
>         >>>>>
>         >>>>> ERROR: No track information for source
>         >>>>> 'AK_assembly_datastore/22/D4/AK_contig_1/'
>         >>>>>
>         >>>>> Looking at the perl script, the file locations are getting read, but
>         >>>>> then the tracks aren't getting found in the hash of commands.
>         >>>>>
>         >>>>> Any advice for getting past this will be most appreciated.  Thank
>         >>>>>you!
>         >>>>>
>         >>>>> Best,
>         >>>>> Val
>         >>>>> _______________________________________________ maker-devel mailing
>         >>>>>list
>         >>>>>maker-devel at box290.bluehost.com
>         <mailto:maker-devel at box290.bluehost.com>
>         >>>>><mailto:maker-devel at box290.bluehost.com
>         <mailto:maker-devel at box290.bluehost.com>>
>         >>>>>
>         >>>>>http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.or
>         >>>>>g
>         >>>>>
>         >>>>>
>         >>>>> _______________________________________________
>         >>>>> maker-devel mailing list
>         >>>>>maker-devel at box290.bluehost.com
>         <mailto:maker-devel at box290.bluehost.com>
>         >>>>>
>         >>>>>http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.or
>         >>>>>g
>         >>>>>
>         >>>
>         >>>
>         >>>
>         >
>         >
>
>
>




More information about the maker-devel mailing list