soda.py

Python-based soda-plot gallery generator

• Description
• Dependencies
• Usage
• Options
• Credits

Description

soda.py is a Python script that generates a gallery of images made from snapshots from a UCSC genome browser instance, so-called "soda plots". Snapshots could be derived from the Altius internal browser instance gb1, or any other UCSC browser instance, if specified.

You provide the script with four required parameters:

1. A BED-formatted file containing your regions of interest.
2. The genome build name, such as hg19, hg38, mm10, etc.
3. The session ID from your genome browser session, which specifies the browser tracks you want to visualize, as well as other visual display parameters that are specific to your session.
4. Where you want to store the gallery end-product.

If the BED file contains a fourth column (commonly used to store the name of the region), its values are used as labels for each page in the gallery.

Additional options are available; please see the Options section.

Note

The BED file does not need to be in BEDOPS sort-bed order. In fact, it can be useful to order the regions in a BED file by some criteria other than genomic position, such as some numerical value stored in the BED file's score column, e.g.:

$ sort -k5,5n input.bed > input_sorted_by_scores.bed

Any ordering is allowed. Gallery snapshots are presented in the same order as rows in the input BED file.

Installation

You can install this from the HPCA module system or clone the code locally. We recommend using the module system.

Modules

From an Altius host which uses the HPCA module system (e.g., sched0.altiusinstitute.org or by way of a job submission via squeue), add the following module to satisfy dependencies, before running soda.py:

$ module add soda

Adding the soda module adds anaconda and ImageMagick modules. Both of these modules are required to run soda.py from an HPCA host.

Download

To grab this kit and run it locally, you can clone it from Github:

$ git clone https://github.com/Altius/soda.git

Note

If you run this script via the Git repository, such from a non-modules host in the lab, or on your personal workstation or laptop, you may need to use pip or similar to add Python libraries that may not be part of a typical environment:

• requests
• requests-kerberos
• Beautiful Soup
• Jinja2
• pdfrw
• pdfminer

It may be useful to use a package manager to add ImageMagick and Ghostscript, if not already installed.

Installing these dependencies may also require administrator privileges. In this case, please see the documentation for these components for installation instructions, or contact your local IT support for assistance.

Usage

As a usage example, you may have a BED file in some home directory called /home/abc/regions.bed. You have a session ID for the gb1 Altius genome browser instance called 123456_abcdef, with all your tracks selected and display parameters set, using hg38 as the reference genome build. Finally, you want to store the results in a folder called /home/abc/public_html/my-soda-plot-results where it can be shared publicly with others via a web browser. Here is a sample command that generates the desired gallery:

$ /path/to/soda.py -r "/home/abc/regions.bed" -b "hg38" -s "123456_abcdef" -o "/home/abc/public_html/my-soda-plot-results"

When finished, the final product is a gallery that is available from the web address http://www.uwencode.org/~abc/my-soda-plot-results.

For example, here is a gallery made from earlier testing that is stored in my home directory's public_html folder: http://uwencode.org/~areynolds/soda-test/

If you run this locally, you can open the result folder's index.html file with your web browser to load the gallery. For example, from the Terminal application in OS X, you can run:

$ open /Users/abc/my-soda-plot-results/index.html

which opens the gallery index in the default web browser.

Options

Required

Four options are required. At minimum:

-r, --regionsFn

Use -r or --regionsFn to specify the path to the input BED file containing regions of interest.

-b, --browserBuildID

The -b or --browserBuildID option specifies the genome build, e.g., hg19, mm10, etc.

-s, --browserSessionID

The -s or --browserSessionID option specify the browser session ID, which references a configuration of tracks and display parameters from a genome browser instance.

-o, --outputDir

Use the -o or --outputDir option to specify where the image gallery is saved. If this path already exists, soda.py will exit with a fatal error message.

Optional

Other options are available depending on how you want to customize the run.

-t, --title

Use -t or --title to specify a gallery title.

[ -i, --addIntervalAnnotation | -d, --addMidpointAnnotation ]

Use -i or --addIntervalAnnotation to add a rectangle underneath all tracks that demarcates the original genomic range (useful when used with --range). Alternatively, use -d or --addMidpointAnnotation to add a vertical line underneath all tracks, centered on the midpoint of the input genomic range. In both cases, the annotation is labeled with the genomic coordinates of the original interval or the calculated midpoint, respectively. It is not allowed to specify both options together.

-w, --annotationRgba
-z, --annotationFontPointSize
-f, --annotationFontFamily
-e, --annotationResolution

When used with -i or -d to add an interval or midpoint annotation, these options may be used to override the default rgba() color, typeface point size, and typeface family (where supported by the local installation of ImageMagick), which are parameters used to render the appearance of the annotation components. The default color is rgba(255, 0, 0, 0.333) and the default point size and font family values are 5.5 and Helvetica, respectively. To keep the file size of the final PDF reasonable, a default resolution for the annotation is set at 150 dpi. This value can be increased, if desired, by overriding this setting. A higher resolution will require more processing time and create larger files.

-j, --outputPngResolution

To reduce file sizes and improve gallery generation speed, PNG-formatted bitmaps are made from the PDF-formatted browser shots with a default resolution of 150 dpi. This value can be overridden with a custom resolution setting to create lower- or higher-quality bitmaps. Note that a higher resolution setting (e.g., 300 or 600 dpi) will require more processing time and the final gallery product will use more disk space.

-a, --range

Use the -a or --range option to pad the BED input's midpoints symmetrically by the specified number of bases. This works regardless of the sort order of the input.

-g, --browserURL

Use the -g or --browserURL option to specify a different genome browser URL other than the Altius gb1 host. If a different host is specified and credentials are required, please use the -u and -p options (see below).

-u, --browserUsername
-p, --browserPassword

Use these two options to specify a username and password for the browser instance, if you pick a different --browserURL and that browser instance requires credentials. If these options are not specified, no credentials are passed along. If authentication fails, soda.py may exit with an error.

-y, --useKerberosAuthentication

Use this option if access to your custom browser instance requires a Kerberos ticket (obtained via kinit, for example).

-v, --verbose

Use -v or --verbose to print debug messages, which may be useful for automation or debugging.

Credits

The general "soda" gallery tool has been authored in various bash- and Perl-flavored incarnations since 2008 by primary authors Richard Sandstrom and Scott Kuehn, with modifications over time by Bob Thurman, Jay Hesselberth, Richard Humbert, Brady Miller and Alex Reynolds.

This Python rewrite and new functionality were authored by Alex Reynolds.

This tool uses the blueimp Gallery and Github octicons projects, both of which are MIT-licensed.

So what's up with the name?

"Scott Kuehn: he came up with the name. The legend is that when asked what to call the program, he lifted a can of cola, sighed deeply, and said: 'soda plot'." - R. Sandstrom