The OpenHelix Blog

Recently while watching the #bioinformatics tag on Twitter I saw Khader Shameer mention Caleydo.  I was instantly hooked at the very clever visualization strategy that they are using to provide more surface area for examining the data you are interested in viewing.  Their specific topics are pathways and gene expression, but it got me thinking about various data types that I would like to see connected in this way.

To skip right over to Caleydo and start trying it out, go here: http://www.caleydo.org/

Caleydo delivers a 3D representation of the expression and pathway data.  The main user interface has an area that is a box.  They call it a bucket, but in my head buckets are round, so I think of this as a box.  On the floor of the box you have a graphic.  But because you also have 4 interior surfaces of the box you have 4 more places to display and link the data.  You can have a heat map microarray representation on one side, and various pathways associated with the genes in that microarray on the other sides.

There’s a short systems biology Application Note in Bioinformatics that describes the framework and gives an overview of the tool.  But there’s also a more detailed paper over at their publication site that will get you started (that 2010 paper for the Visualization conference in Taipei).

My computer is a bit underpowered, but I was able to load their webstart version and begin to look around.  They provide some sample data you can select and examine.  For the movie this week, though, I was unable to load that and run the recording software at the same time.  So mostly it’s an introduction to the concept and the site.  You’ll have to go over and load it up yourself to try it out.  If the webstart version doesn’t work for you, there are a couple of other download options for different platforms.

The Caleydo team has also done a YouTube overview of the features that you can examine.

http://www.caleydo.org

So try out this visualization strategy and see what you think.  I really like the concept.

+++++++++++++++++++++

Streit, M., Lex, A., Kalkusch, M., Zatloukal, K., & Schmalstieg, D. (2009). Caleydo: connecting pathways and gene expression Bioinformatics, 25 (20), 2760-2761 DOI: 10.1093/bioinformatics/btp432

So far at OpenHelix we’ve generally been focusing on the more broad tools that are of use to the largest cross-section of biomedical researchers.  Everybody needs genome browsers pretty much at some point in their research.  However, I’ve always had a soft spot for tissue-specific resources.  Since my PhD project was on muscle, I have a lot of thoughts on tissue-specific regulation, expression, and splicing that I think are going to be just fascinating as we build on the whole genome sequencing base projects.  A lot of the “hypothetical” and “unknown” predicted sequences are going to fall out of spatial-and temporal-specific expression projects as we move forward.

That appears to be the case in liver according to a new proteomics paper.  There’s a commentary in the Journal of Proteome Research that speaks to this:

Mammoth data set from human liver reported by Quinn Martin Eastman

A total of 6788 proteins were identified, though that number excludes >6000 proteins that had only one peptide match and were eliminated from the final count. The researchers identified proteins corresponding to 60% of all of the protein-encoding genes expressed in liver, which were identified by RNA analysis from the same samples….

Some 3721 of the identified proteins had not been seen in human liver before, though they had been detected in other human organs. Almost 1000 were “hypothetical”—their existence previously inferred from DNA sequence information only.

Emphasis mine. That’s very cool stuff.

The commentary refers to a paper in the same issue from a consortium of researchers, and links to the resulting database for the project.  The database has about the most complicated name I’ve ever seen, visible in the paper title here:

First Insight into the Human Liver Proteome from PROTEOME^SKY-LIVER^Hu 1.0, a Publicly Available Database

Hyphens and superscripts–oh my!

You can check out the corresponding databases associated with this project at 2 URLs:

dbLEP for Liver Expression Profile:  http://dblep.hupo.org.cn

and Liverbase: http://liverbase.hupo.org.cn

I’m looking forward to other tissue collections as well.

In today’s tip, I want to show you how to use a great looking beta tool that I just found at EBI’s ArrayExpress Gene Expression repository (AE). The tool’s name is the ArrayExpress Atlas. You may have retrieved expression data from the ArrayExpress Warehouse, which is a carefully curated collection of expression data. The Warehouse is a wonderful resource, and a great way to obtain expression data sets, but the information retrieved is organized by gene name and sample values. The ArrayExpress Atlas appears to be the next generation of the Warehouse and it provides gene expression data as a table, with genes corresponding to rows and experimental conditions corresponding to columns. The tool is easy to use, provides easy to interpret results, and looks like its capabilities are growing fast. Check out this tip, check out the Atlas blog spot, check out the tool, and send any feedback for improving the tool to AE. 

There are times when being able to download the data used to create a publication is very useful. Perhaps you want to compare interesting data to some of your own. Or you may want to analyze the data differently than the original author to try and gain additional information. Or perhaps you’d just like to check out the data yourself that the author’s conclusions are valid. For my “tip of the week,” I thought I’d give you a tip for downloading some of that data from NCBI’s GEO (or Gene Expression Omnibus) database. We’ve got a whole introductory tutorial on how to use GEO, but this ~3 minute screencast will show you how to easily get GEO information into an Excel spreadsheet. Hope you find this tip as useful as I did when I learned it!