Wednesday, April 9, 2014

v37: More options and bug fixes

If you're upgrading from previous version, you need to re-compile "subtleKnife.c". This version has many changes listed below:

1. App finder. Go to Apps and use the text box to find apps:

You may be interested to look for these hidden apps: Metadata term finder, Sequence fetcher, Datahub validator.

2. Validate datahub. Search by "hub" or "validate" to find this app:

This is a preliminary version of hub validator, and is capable of validating tracks (if there's no JSON errors anywhere in the hub). It won't add or register custom tracks in the hub. It is expected the hub to be of same reference genome as the one the browser is running on.

3. Better handling of broken tracks. Previously only numerical tracks can be handled.

4. Correct SVG color coding, solved an error when there's transparency involved. Thanks to Kamil Slowikowski!

5. CHG methylation. In methylC track, when you combine cytosine methylation level from two strands, only CG cytosines are combined, but not CHG cytosines since they are highly asymmetrical in mammalian genomes. But now we give user the option to combine that, by checking an additional checkbox:

Combining CHG (CAG/CTG) methylation data is desirable in certain cases (strong symmetry of CHG methylation in the plant genomes). This option will only appear when there's strand-specific CHG data present in a methylC track. We thank our anonymous reviewer for helpful discussions and suggestion of this feature.

Friday, April 4, 2014

v36: upload hammock-format file

Version 36 features numerous improvements.

Upload hammock file

You now have more convenient options to display hammock tracks on the browser.

  1. Prepare your hammock track, host it on your web server, and go to Custom track panel to submit it.
  2. Or, simply upload the hammock-format text file and the browser will display the track.

When you are either loading hammock track via URL or file uploading, you can find a text box to enter JSON descriptions. You can always refer to the example track by clicking the "Use example" button:

Peak-calling data

We keeps enhancing the hammock format. As a result, 3 peak-calling formats (narrowPeak, broadPeak, gappedPeak) can be converted and displayed as hammock tracks. We also provide scripts to convert peak to hammock formats. Please refer to our wiki page for more information:

Bar chart background

We implemented this feature a while ago. We've just made it easier to use, by adding an option in the configuration panel. By using it, spots with low values can be visually separated from spots without any data (only applicable for bedGraph tracks).

Shared Y axis scale

A common Y scale range can be shared among a group of numerical tracks, through the "group" attribute. This can be achieved by using the "group" attribute when defining a track in a datahub. In future we will implement UI for you to conveniently choose and assign tracks to groups.

Track height and Y scale are kept the same between main/split browser panels

The split panel function is now more appealing by tightening up this loose end!

Saturday, March 22, 2014

v35: hammock tracks for richly annotated genomic features

The hammock tracks is a new track type for displaying richly annotated genomic feature on the browser. It's simple and straight, succinct by having only 4 columns, versatile for the JSON field, and can do awesome things. The current format specification is here:

The format is under active development, and suggestions are always welcome. It will also remain backward-compatible by its JSON nature.

bgzip and tabix should be used to compress and index a hammock file for browser display, same as the bedGraph track.

Usually you need to submit a hammock track via a JSON datahub, as can be seen in this example:

Show this example.

A few screenshots:

Gencode track. Featuring the structure display for gene models, and color assignments for different types of genes (coding/non-coding/pseudo...):

Two classes of repeat features displayed in one track (SINE and DNA). Degree of darkness is determined by the Swiss-Waterman score of each feature (with regard to max score of all features in the view range):

The same repeat track as above, displayed in "barplot" mode, showing the Smith-Waterman scores displayed as bars on top of repeat features. Bar width is the same as the on-screen width of that item, and bar color is similar but slightly lighter than the item color:

Another barplot example showing exon array data. featuring the overlapping display of bars (for items that overlap with each other):

 NarrowPeak data display, with the tooltip showing detailed information on the clicked item: