release-notes.txt

{{$NEXT}}

 * Added a new direct-access storage driver for BAM files, removing
   the need for `bam-to-json.pl`.  This new method of BAM access is
   far superior to the old `bam-to-json.pl` in nearly every way,
   except in browser compatibility.  Like the BigWig direct access
   backend added in JBrowse 1.5.0, it is based on code from Thomas
   Down's Dalliance Genome Explorer, and works in all major browsers
   *except* Internet Explorer, because IE lacks support for the
   necessary web standards.  It may work with Internet Explorer 10,
   but this has not been tested yet.

 * Added a new `Alignments` track type designed to work seamlessly
   with BAM files.  This track type shows basepair differences and
   insertions between aligned reads and the reference, and highlights
   reads with missing mate-pairs in red.

 * Added the ability to export track data in FASTA, GFF3, bed,
   bedGraph, and Wiggle formats (issue #104).  To export data, turn on
   the track of interest, then click on its track label to bring up
   the track menu, and select "Save track data".

 * Added a new `Wiggle/Density` track type, analagous to the GBrowse
   `wiggle_density` glyph type.  Shows the Wiggle information using
   varying intensity of color (issue #66).  Renamed the `Wiggle` track
   to `Wiggle/XYPlot`, and made the old `Wiggle` track type an alias
   to `Wiggle/XYPlot`.

 * Both `Wiggle/XYPlot` and `Wiggle/Density` now support a `style ->
   bg_color` option.  Color-density plots blend the `pos_color` or
   `neg_color` into the `bg_color` in amounts that vary with the
   wiggle data, and xyplots fill the background color behind all
   points that have data present, regardless of value.  `bg_color`
   defaults to off for xy xplots, and semi-transparent gray for
   density plots.  Setting this makes it easier for users to
   distinguish at a glance between regions with no data, and regions
   with a value of 0.

 * Added a new `FeatureCoverage` track type, which shows a
   dynamically-computed XY-plot of the depth of coverage of features
   across a genome.  One good use of this track type is to provide a
   quick coverage plot directly from a BAM file.  However, since this
   track calculates coverage on the fly, it can be slow when used with
   large regions or very deep coverage.  In this case, it is
   recommended to generate a BigWig file containing the coverage data,
   and display it with a `Wiggle/XYPlot` or `Wiggle/Density` track.

 * DNA bases are now displayed with color-coded backgrounds, allowing
   basepair information to be discerned when zoomed somewhat further
   out, when base letter cannot be drawn.  Colors are also designed to
   match the base-mismatch colors used in `Alignment` tracks, enabling
   clearer SNP visualization.

 * Added a vertical line cursor and labels showing the current
   basepair position of the mouse when hovering over the scale bar, or
   when doing a rubber-band zoom (Erik Derohanian) (issue #32).

 * Added an animation to make it easier to see where in the track
   selection list a closed track has gone, when using the simple track
   selector (issue #151).

 * Information dialog boxes are now easier to dismiss: clicking
   anywhere outside of them, or pressing any key, will make them go
   away.

 * Improvements to feature track configuration:
   * Feature tracks no longer use the `style.subfeatureScale`
     configuration variable to determine whether to show subfeatures.
     Instead, subfeatures are shown if the parent feature, when shown on
     the screen, is wider than `style.minSubfeatureWidth`, which
     defaults to 6 pixels.
   * Make explicitly-configured track `labelScale`, `histScale`, and
     `descriptionScale` not be modulated by the feature density: only
     use the feature density to pick the scale defaults

 * The default feature-detail dialog box now shows more information, adding:
   * the feature's exact length
   * full details of its subfeatures

 * Added a `locationBoxLength` configuration variable that controls
   the width of the location box.  In addition, the default width of
   the location box is now also smarter.  Instead of a fixed 25
   characters, it is calculated to fit the largest location string
   that is likely to be produced, based on the length of the reference
   sequences and the length of their names.

 * Pressing SHIFT+-up/down arrow keys now cause the genome view to
   zoom in and out.  If ALT is added, it zooms further.  Thanks to
   Karsten Hokamp for the excellent suggestion.

 * Holding SHIFT while scrolling left and right with the arrow keys
   causes the view to scroll further.

 * Added a `theme` configuration variable to allow changing the
   graphical theme to something different from the default "tundra".
   However, no other themes are implemented yet.

 * Greatly sped up rendering of HTML subfeatures by caching the
   heights of subfeature HTML elements.

 * Fixed bug in which the genome view executed a double-click zoom when
   users rapidly clicked on multiple track 'close' buttons.

 * Fixed bug with the genome view scrolling in response to arrow keys
   being pressed when typing in the location box.

 * Fixed bug in which the score display in Wiggle tracks would
   sometimes flicker when moving the mouse.

1.6.5     2012-10-26 12:10:08 America/New_York

 * The location box now shows the length of the currently visible
   region in parentheses.  For example: `ctgB:1244..3566 (2.32 Kb)`.
   Thanks to Karsten Hokamp for the suggestion!

 * The arrow keys on the keyboard can now be used to pan and scroll
   around the genome view.

 * Wiggle track mouseover cursors now display the score with only 6
   significant digits, avoiding confusion over approximations
   introduced by scores being converted to IEEE floating-point numbers
   (as for BigWig files) and back to text.

 * The faceted track selector now renders any HTML that may be present
   in the track metadata by default.  To turn this off, it now accepts
   an `escapeHTMLInData` option that, if set to `true` or `1`, will
   not render the HTML, but will instead display the raw code
   (issue #145).

 * Upgraded to a more recent version of jszlib, which contains some
   important bugfixes (issue #157).

 * Fixed bug dealing with very large regions with the same value in
   canvas-based Wiggle tracks (also issue #157).

 * `prepare-refseqs.pl` now by default uses a more scalable directory
   structure for storing sequences.  This fixes problems some users
   were experiencing with large numbers of reference sequences
   (issue #139).

 * `ucsc-to-json.pl` now supports a `--primaryName` option allowing
   the users to alter which UCSC data field is displayed by JBrowse as
   the primary name of the features in a track.  Also,
   `ucsc-to-json.pl` now treats as indexable names all UCSC data
   columns called "name", "alias", or "id" followed by zero or more
   digits.  Thanks to Steffi Geisen for this suggestion.

 * Fixed "Duplicate specification" warnings coming from
   `flatfile-to-json.pl` and `remove-track.pl`.

 * Fixed bugs in which both the first few and the last few bases of a
   reference sequence were not displaying correctly in a DNA track.

1.6.4     2012-10-16 11:50:44 America/New_York

 * Improvements to the scalability of `generate-names.pl`.  Many
   thanks to Steffi Geisen for her ongoing help with this.
    * Users can now manually specify which tracks will be indexed to
      enable autocompletion and searching for their feature names in
      the JBrowse location box.
    * The lazy-trie name indexing structure now correctly handles the
      case of large numbers of features that may share the same
      name. Before, it was generating files that were too large for
      the client to handle.

 * Fixed off-by-one error in Wiggle track display code: wiggle data
   was incorrectly displayed shifted one base to the left of its
   proper position.  Thanks to Steffi Geisen for noticing this.

 * Fixed bug in which the reference-sequence selection box did not
   automatically update in all situations to reflect the current
   reference sequence.

1.6.3     2012-09-28 11:33:36 America/New_York

 * Fixed bug with shift-rubberband-zooming not working on Windows
   (issue #150) (Erik Derohanian).

 * Fixed "Can't locate JSON.pm" errors with add-track-json.pl.

 * Added a reference-sequence-selection dropdown box, similar to the
   old one, that is on by default if fewer than 30 reference
   sequences, otherwise it's off unless `refSeqDropdown: true` is set
   in the configuration (issue #138).

 * Fixed bug in which popup dialog boxes showing other websites showed
   the website in only the top portion of the dialog box.  Only
   present in some browsers (issue #149).

 * Fix coordinate display bug in feature detail popups.  The feature's
   position was being displayed in interbase coordinates, but should
   be displayed in 1-based coordinates.  Thanks to Steffi Geisen for
   pointing this out.

 * Added a `style.height` option to Wiggle tracks to control the
   track's height in pixels (issue #131) (Erik Derohanian).

 * Added support for a `style.trackLabelCss` configuration variable to
   allow customizing the appearance of the label for a particular
   track (issue #4) (Erik Derohanian).

1.6.2     2012-09-10 17:47:43 America/New_York

 * Fixed feature-layout performance problem when zoomed very far in
   on features that are much larger than the viewing window.

 * Added a default `menuTemplate` to all HTML-based features, so that
   all HTML features now have a right-click menu by default.

 * Add `css` configuration variable that allows users to specify
   either strings or URLs containing CSS to add.

 * improved `bin/ucsc-to-json.pl` error messages

 * `bin/add-track-json.pl` now replaces tracks in the target
   configuration if they have the same label.

1.6.1     2012-08-28 16:50:41 America/New_York

 * JBrowse now attempts to prevent feature labels being obscured by
   track labels by keeping the feature labels a bit further away from
   the left side of the view, if possible.

 * Fixed bug in which welcome page was not shown when JBrowse has not
   been configured yet (issue #130).

 * Fixed bug in which passing `tracklist=0` (as when running the
   browser in embedded mode) caused JBrowse to crash (issue #132).

 * Added dependency on Bio::GFF3::LowLevel::Parser 1.4, which has an
   important bugfix related to multi-location features (issue #109).
   Thanks to Victor Poten for help in isolating this problem.

1.6.0     2012-08-25 18:04:44 America/New_York

 * Added `description` capabilities to HTML-based features, similar to
   GBrowse's descriptions.  If zoomed in far enough (as defined by
   `style.descriptionScale`), adds a second line of text below the
   feature's label that shows the content's of the featur's `Note` or
   `description` attribute (issue #67).

 * Give prepare-refseqs.pl the capability to load reference sequences
   from embedded FASTA sections in GFF3 files (issue #128).

 * Configuration files can now recursively `include` eachother, and
   `tracks` sections are merged intelligently.

 * Made sequence tracks not disappear when zoomed out too far to see
   base pairs.  Instead, sequence tracks simply display lines
   suggesting that DNA would be visible at higher
   magnification. (issue #124).

 * Double-clicking track labels in the simple track selector now turns
   the track on (issue #123).

 * Fixed bug in BigWig tracks that use the "scale": "log" option: did
   not render properly when the wiggle data included 0's. This was due
   to the fact that the origin was being mapped to Infinity. Thanks to
   the [Mockler Lab](http://www.mocklerlab.org/) for the fix! (issue
   #127).

 * Fixed bug in NCList binary search code in which zero-length
   features at node boundaries would not be found (fix by Ed Lee).

 * Fixed bug in which dragging the scroll bar on the simple track
   selector can sometimes cause a drag-and-drop to erroneously begin
   (issue #89).

 * Fixed some bugs in the layout of HTML-based features in which
   features in different blocks would overlap in some circumstances.
   Under the hood, replaced the contour-based layout engine with a
   simpler, not-much-slower implementation that is more correct in the
   general case (issue #122).

 * Fixed a bug with vertical centering of strand arrows and other
   sub-elements of HTML-based features.

1.5.0     2012-08-13 15:37:27 America/New_York

 * Added a direct-access storage driver for BigWig data files, based
   on code from the Dalliance Genome Explorer by Thomas Down.  BigWig
   file access is supported now by the current versions of all major
   browsers except Internet Explorer (which is expected to work when
   version 10 is released along with Windows 8).

 * Added a `canvas`-based wiggle track implementation for quantitative
   data that, when used with the new BigWig storage backend, removes
   the need to pre-generate rendered images of wiggle data.  Its
   display is also highly configurable, with configuration options
   modeled on the GBrowse `wiggle_xyplot` glyph type
   (i.e. `Bio::Graphics::Glyph::wiggle_xyplot`).

 * Added highly configurable behavior for left-clicking and
   right-clicking features in HTML-based feature tracks.  If a
   `menuTemplate` option is specified in the track configuration,
   right-clicking a feature brings up a context menu, the items in
   which can be configured to do nearly anything, but that are easy to
   configure for the very common use case of wanting to display
   content from a certain URL.  Feature left-clicks are also
   configurable using the same mechanism.  Thanks to Alexie
   Papanicolaou and Temi Varghese for the initial implementation of
   context menus.

 * Improved the default HTML feature left-click dialog box.  It is now
   both prettier, and more comprehensive, displaying all available
   data for the feature.

 * Added the long-missing ability to render a second line of label
   text for features containing their description, which is taken from
   the feature's 'Note' or 'description' attribute, if present.  The
   description is only displayed if the track's `description`
   configuration variable is set to a true value, which is the
   default.  There is also a style.maxDescriptionLength value, that
   can be set to control how long a description can be before it is
   truncated with ellipses.  It defaults to 70 characters.

 * Added a small helper script, `add-track-json.pl` that developers
   and advanced users can use to programmatically add a block of track
   configuration JSON to an existing JBrowse configuration file.

 * Improved / fixed vertical alignment of sub-elements of HTML
   features, including subfeatures and the arrowheads that show
   strand.  All elements in a feature are now vertically centered by
   default.

1.4.2     2012-07-12 15:38:55 America/New_York

 * Restore support for histScale, subfeatureScale, and labelScale in
   --clientConfig and track configuration JSON.  Thanks to Hernán
   Bondino for pointing this out.

1.4.1     2012-07-10 14:58:34 America/New_York

 * Made displayColumns option for the faceted track selector
   case-insensitive, and interpret a column name of "Name" as meaning
   the track's key.

 * `bam-to-json.pl` now filters out alignments that are not at least
   two nucleotides in length.  Thanks to Tristan Lubinski for
   assistance.

 * Introduced limits on the sizes of cookies that can be set,
   preventing '400 bad request' errors (issue #113).

1.4.0     2012-06-14 17:43:50 America/New_York

 * Added a full-featured faceted track selector for users that have
   many (hundreds or thousands) of tracks.  This can be turned on by
   setting the `trackSelectorType` config variable to "Faceted".  See
   the JBrowse wiki for more documentation on how to use faceted track
   selection. (issue #95)

 * Removed the dropdown selector for reference sequences in favor of
   making the location box auto-complete reference sequence and
   feature names.  This makes JBrowse much more scalable to large
   numbers of reference sequences. (fixes issues #3, #60, and #101)

 * Added a vertical-scrolling marker on the right side of the track
   pane, making it much easier to discern the vertical position of the
   track display. (issue #93).

 * biodb-to-json.pl and flatfile-to-json.pl now load all available
   feature data: all attributes of features are now encoded in the
   JSON and are available for use by feature callbacks. (issue #72)

 * Feature labels now do not scroll off screen if any part of the
   feature is still visible (fixes issue #62).

 * Added jbrowse_conf.json, a default JSON-format configuration file,
   to the JBrowse root directory.  Makes it easier to get started
   with more advanced JBrowse configuration.

 * JBrowse instances now report usage statistics to the JBrowse
   developers.  This data is very important to the JBrowse project,
   since it is used to make the case to grant agencies for continuing
   to fund JBrowse development.  No research data is transmitted, the
   data collected is limited to standard Google Analytics, along with
   a count of how many tracks the JBrowse instance has, how many
   reference sequences are present, their average length, and what
   types of tracks (wiggle, feature, etc) are present.  Users can
   disable usage statistics by setting "suppressUsageStatistics: true"
   in the JBrowse configuration.

1.3.1     2012-04-19 17:55:44 America/New_York

 * Fixed memory-management bug that caused way too much RAM to be used
   by FeatureTrack loading (flatfile, bam, and biodb-to-json.pl) when
   loading with very large numbers of reference sequences.  Big thanks
   to Tristan Lubinski for help in isolating this.

 * Fixed some bugs in BAM support section of setup.sh autosetup
   script, thanks to Tristan Lubinski for help in isolating this as
   well.

 * Added an example document with an iframe running JBrowse in
   embedded mode in docs/examples/embedded_mode.html

 * flatfile-to-json.pl now loads the 'score' attribute of features in the
   JSON.

1.3.0     2012-04-13 17:04:30 America/New_York

 * Added support for "rubberband" dynamic zooming, in which users can
   click and drag to select a region to zoom to.  Dragging on any
   scale bar, or shift-dragging on the main track pane, triggers a
   dynamic zoom.

 * Correcting a long-standing oversight, wiggle data tracks and
   feature histograms now have numerical y-axis scales that show the
   numerical values of the data.

 * Server-side data-formatting scripts now support a --compress option
   to compress (gzip) feature and sequence data to conserve server
   disk space.  Using this option requires some web server
   configuration.  Under Apache, AllowOverride FileInfo (or
   AllowOverride All) must be set for the JBrowse data directories in
   order to use the included .htaccess files, and mod_headers and
   mod_setenvif must be installed and enabled.  Under nginx a
   configuration snippet like the following should be included in the
   configuration:

      location ~* "\.(json|txt)z$" {
               add_header Content-Encoding  gzip;
               gzip off;
               types { application/json jsonz; }
      }

 * flatfile-to-json.pl: now much faster and more memory-efficient,
   especially for GFF3 files.  Remember that '###' directives are very
   important to have in large GFF3 files!  Also removed nonfunctional
   --extraData switch.

 * Added ability to turn off some JBrowse UI panels via URL arguments
   to the default index.html, or via arguments to the Browser
   constructor itself.  Can dynamically turn off the navigation box,
   the overview panel, and the track list, respectively.  When all of
   these are off, and if run in an iframe, JBrowse is running in an
   "embedded mode" that looks similar to the output of GBrowse's
   gbrowse_img script, with the exception that the view in this case
   is a fully functioning, scrollable and zoomable JBrowse.  See the
   GMOD wiki (http://gmod.org/wiki/JBrowse) for more on how to set up
   embedded mode.  Thanks to Julie Moon, a co-op student working at
   OICR, for this work!

 * Improved graphical look and feel.

 * Browser support for this release:
      * Google Chrome 18                perfect
      * Google Chrome 17                perfect
      * Mozilla Firefox 11.0            perfect
      * Mozilla Firefox 10.1            perfect
      * Mozilla Firefox 10.0.2          perfect
      * Mozilla Firefox 3.6.28          nonfunctional
      * Apple Safari 5.1.5 (Lion)       perfect
      * Microsoft Internet Explorer 9   perfect
      * Microsoft Internet Explorer 8   good
      * Microsoft Internet Explorer 7   minor problems
      * Microsoft Internet Explorer 6   not tested
      * KDE Konqueror 4.7.4             nonfunctional
      * KDE Konqueror 4.5.5             nonfunctional
      * Opera (all versions)            not tested

   NOTE: Internet Explorer 6 is no longer supported by JBrowse.

 * Added an automated-setup script, setup.sh, that tries to install
   Perl prerequisites, format Volvox example data, and install Wiggle
   and BAM support (fetching samtools from SVN if necessary)
   automatically.

 * Navigating to JBrowse with missing or malformed configuration or
   data will now bring up an error page with useful messages and links
   to help documentation, instead of a blank white page.

 * JBrowse data directories now include an Apache .htaccess in their
   root directory that, if mod_headers is installed and AllowOverride
   FileInfo or AllowOverride All is enabled, will emit the proper HTTP
   headers to allow cross-origin XHR requests for the data.

 * A new "Help" link in the upper right, or pressing "?" on the
   keyboard, brings up a "JBrowse Help" dialog box with basic usage
   information and links to more help information.

 * Arrowheads indicating strandedness are now drawn inside feature
   boundaries.

 * Clicking on the overview bar or the main scale bar now centers the
   view at the clicked position.  In addition, while holding down
   shift, clicking in the main track panel will also center the view
   at that position.

 * Added bin/remove-track.pl, a script to remove a track from a
   JBrowse data directory.  Run bin/remove-track.pl -? to see its
   documentation.

 * Added build instrumentation to support a JSDoc-based system of
   developer API documentation.  This documentation is still far from
   complete.

 * Ian Davis contributed code to add a view of the reverse strand of
   the sequence in the DNA track.  Thanks Ian!

 * Fixed bug in which, for some sequence chunk sizes, the DNA bases
   would display incorrectly.

 * Added minor gridlines to the main track view.

 * Fixed a long-standing off-by-one bug where the window could not be
   scrolled to view the last base in the reference sequence.

 * Coordinates displayed in the user interface are now 1-based closed
   coordinates, which are more familiar to most users.  Previously,
   the labels displayed interbase (i.e. 0-based half-open)
   coordinates.

 * NON-BACKWARDS-COMPATIBLE improvements to the JSON format used for
   track configuration, feature data, and image data
     - initial support for a new hook system for greater
       administrator-configurability of feature display
     - support for more than one level of subfeatures

 * Miscellaneous improvements and refactoring of data-formatting Perl
   code.

 * More detailed POD-based help documentation on all scripts

 * --tracklabel options to all scripts replaced with --trackLabel.

 * New suite of integration tests, and some unit tests, for
   server-side Perl code.

 * Beginnings of a suite of Selenium-based integration tests for the
   front-end JavaScript code.

 * Support for Apple touch-based devices merged into normal
   index.html, so that the same link can be used regardless of the
   browsing platform.

 * Bug fixed in which non-stranded features do not display properly
   (a problem with the CSS styles).

========================================================

version 1.2.1, March 2011

Fixed bugs in release 1.2:

- problems with BED handling reported by Gregg Helt and Brenton Graveley

- performance regression reported by Chris Childers

- incorrect handling of names in ucsc-to-json.pl

- wig2png slowdown reported by Brenton Graveley

========================================================

version 1.2, Febrary 2011

These notes document changes since release 1.1 in September 2010.

Most of the work in this release went into making JBrowse handle large
amounts of feature data better.  Before, the amount of memory used
when processing BAM files was more than 10 times the size of the file;
now, the amount of memory required is fixed.

Other new features in this release:

- Import of UCSC database dumps.  A ucsc-to-json.pl script is now
  provided for taking database dumps from UCSC and creating a JBrowse
  instance using them.  The "genePred" and "bed" track types are
  currently supported; "psl" tracks are not yet supported.

- Touch.  Juan Aguilar's code for using JBrowse on an iOS device
  (iPhone, iPod touch, iPad) is now integrated.  As of the current
  release, users wanting to use JBrowse on those devices have to
  navigate to a separate HTML page (touch.html) rather than the
  default index.html; i.e. the code does not currently detect
  touchscreen devices automatically.

- Bug fixes. A number of bugs have also been fixed, including one that
  restricted the placement of the "data" directory, and a bug in
  wiggle rendering that caused spurious peaks or troughs at tile
  boundaries.

Known issues/limitations with this release:

- Some additional CPAN modules are now required:

    PerlIO::gzip
    Heap::Simple
    Devel::Size

- No JSON-level backward compatibility. If you are upgrading from an
  older version of JBrowse, you will have to regenerate all the JSON
  files on your server. This means wiping your jbrowse/data directory
  and re-running all server scripts (flatfile-to-json, biodb-to-json,
  wig-to-json, etc.) to regenerate your data from the original
  FASTA/GFF/BED/WIG files. We apologize for the inconvenience of this,
  but it is inevitable sometimes; we do aim to minimize the number of
  releases which are backwardly-incompatible in this way.

========================================================

version 1.1, September 2010.

These notes document JBrowse developments in the period from July 2009
(online publication of the first JBrowse paper in Genome Research, the
de facto "version 1.0") up to September 2010 (the first in a planned
series of quarterly releases).

New features in this release:

- Scalability. JBrowse can now handle very large data tracks,
including human EST/SNP tracks, or tracks of next-gen sequence
reads. Large datasets are broken into smaller chunks, so there is no
loading delay for big tracks.

- Extensibility. A Perl module (ImageTrackRenderer.pm) for creating
user-drawn image tracks is now available, based on the CPAN GD.pm
module. An example program is provided, draw-basepair-track.pl, that
uses this module to draw arcs over a sequence representing the
base-pairing interactions of RNA secondary structure.

- Bug fixes. Numerous display glitches have been fixed, including
issues with wide-screen monitors and long mostly-offscreen features.

Known issues/limitations with this release:

- No JSON-level backward compatibility. If you are upgrading from an
older version of JBrowse, you will have to regenerate all the JSON
files on your server. This means wiping your jbrowse/data directory
and re-running all server scripts (flatfile-to-json, biodb-to-json,
wig-to-json, etc.) to regenerate your data from the original
FASTA/GFF/BED/WIG files.

- Next-gen sequence display is currently restricted to the
co-ordinates of the outermost region to which a single read is
mapped. There is no support (yet) for displaying pairing between
reads, sequences of reads, alignment of read to reference sequence
(e.g. splicing), or mismatches between read and reference.

- Processing SAM/BAM next-gen sequence files takes a lot of memory
(about 500 megabytes per million features).

- Numerical (as opposed to comparative) readout of the data in Wiggle
tracks, e.g. via a y-axis label or mouseover popup, is still
unsupported.

Please visit the website for contact info

http://jbrowse.org/