DEIVA: a web application for interactive visual analysis of differential gene express

[anton.kratz]

Dear SEQanswers members,

we have developed a web app, DEIVA, for interactive visual analysis of differential gene expression results, such as you typically get from DESeq2 and edgeR.

The core functionality of DEIVA is the ability to identify and locate genes in density or scatter plot representations in a way that meets user expectations, with minimal effort, virtually no learning curve, and no dependencies on any browser plug-ins or need for the end user to install anything. Almost all processing is done client side, so DEIVA scales very well with increasing numbers of users and datasets. We found it useful for exploring differential gene expression plots in the context of several research projects.

Typical scenarios why you might want to deploy DEIVA:

• Sharing results with others, especially colleagues who are not inclined or do not have the time to use R/Bioconductor, and enabling them to do some analysis in a very immediate manner
• Discussing differential gene expression results in meetings, looking up things "live"

Features

• Identify genes by brushing.
• Locate genes by searching for their name. Multiple genes can be located at the same time (separated by a space), and are automatically highlighted in different colors.
• Search for genes in the data table.
• Zoom: use the toolbar to zoom into the selected rectangle (home button zooms out).
• Filter: two different cut-off sliders which are linked by logical "and". Number of genes passing the filter, up and down, is shown.
• Mobile device support. The user interface adapts to desktop as well as mobile environment devices.
• Just select pre-loaded data from a dropdown, or:
• User data. Drag and drop formatted differential gene expression results files onto the plot area to view them in DEIVA.
• MIT license

A running live instance with data to play around with is at
http://hypercubed.github.io/DEIVA/
No sign up needed!
Th green "Show me!" button in the upper right corner starts a quick tour.

I am posting this here in SEQanswers because we are very interested in getting feedback on DEIVA. Either here on the SEQanswers forum, or in the form of GitHub issues, or by email. I believe and hope some of you might find it useful - please let us know what you think!

The project homepage is
https://github.com/Hypercubed/DEIVA

Thank you,
Anton

[colindaven]

Interesting!

What would you say are the key advantages or disadvantages of DEIVA with respect to DEGUST ? http://www.vicbioinformatics.com/degust/

Is it, say, easier to install (Degust is _very_ hard) ?

[anton.kratz]

Hi colindaven,

DEIVA is extremely easy and fast to install - please have a look at "Deploying DEIVA with your own data" in the GitHub readme. There are no server dependencies or plug-ins needed, just clone the gh-pages repo and replace the example data with your own data. If you just want to quickly look something up, this is even easier, in that case you can just drop your own file onto the "Drop your own data here" button; it is safe, no data is sent anywhere, everything stays client side.

Compared to DEGUST: DEIVA is the only tool that has gene lookup. This is great for looking up lists of marker genes or any gene list of interest. It is also far easier to use than DEGUST in my opinion. The aim behind DEIVA was to make it as easy to use as, say, Google mail.

Pros of DEGUST: it has more functionality than DEIVA. But again, the motivation to create DEIVA was to make a tool for identifying and looking up genes as easily as is possible, so that bioinformaticians can communicate and share their results with domain experts easily while meeting user expectations (easy to understand and use like any other web app, mobile device support), being scalable to very large projects, and secure.

If you want more functionality than what DEIVA provides, I recommend setting up something like Gene Pattern by Broad Institute (http://www.broadinstitute.org/cancer...e/genepattern#), this requires some heavy lifting though. Their Multiplot module is somewhat similar to DEIVA, but again: without gene lookup as fas as I can see.

We are working on a manuscript where we also compare DEIVA to several other tools in a feature matrix.

[colindaven]

Thanks, that's a great answer. I will look at looking DEIVA in the future if it is easy to install since there is a lot of value in this kind of app where a biologists' exploration of their own data is not tied to availability of a bioinformatician.

[anton.kratz]

Originally Posted by colindaven

Thanks, that's a great answer. I will look at looking DEIVA in the future if it is easy to install since there is a lot of value in this kind of app where a biologists' exploration of their own data is not tied to availability of a bioinformatician.

Thanks for the comment. If you want to try it out you don't need to install anything - just can just go the live instance at http://hypercubed.github.io/DEIVA/ and try a preloaded example dataset, no signup or anything needed.

[anton.kratz]

There is now also a publication for DEIVA:

Harshbarger J, Kratz A, Carninci P. 2017. DEIVA: a web application for interactive visual analysis of differential gene expression profiles. BMC Genomics 18. http://bmcgenomics.biomedcentral.com...864-016-3396-5

[Utkarsh Kapoor]

Originally Posted by anton.kratz

Thanks for the comment. If you want to try it out you don't need to install anything - just can just go the live instance at http://hypercubed.github.io/DEIVA/ and try a preloaded example dataset, no signup or anything needed.

Hi Anton,
Is it possible to change the pre-uploaded data in project-chi-electron-app? I would like to upload my deseq2 files here?

I am having troubles installing from the source code and making the changes suggested on the website. I am not a bioinformatician so I would'nt be surprised if I am doing something extremely stupid. If data cannot be changed to the electron-app, I would be grateful for a step by step installation instruction.

Thanks,

Utkarsh

[anton.kratz]

Originally Posted by Utkarsh Kapoor

Hi Anton,
Is it possible to change the pre-uploaded data in project-chi-electron-app? I would like to upload my deseq2 files here?

I am having troubles installing from the source code and making the changes suggested on the website. I am not a bioinformatician so I would'nt be surprised if I am doing something extremely stupid. If data cannot be changed to the electron-app, I would be grateful for a step by step installation instruction.

Thanks,

Utkarsh

Hi Utkarsh,

instead of installing from source, a far easier way is to download the gh-pages branch Zip file from https://github.com/Hypercubed/DEIVA/...e/gh-pages.zip

You can just download this zip-file and unpack it somewhere, and replace the example data with your data, and adjust the index.tsv file.

See section “Deploying DEIVA with your own data” in the readme at https://github.com/Hypercubed/DEIVA#...-your-own-data

For this to work, probably (depends on your browser and maybe other external circumstances) it might be necessary that some web server is running. The reason is because many browsers do not run JavaScript anymore if you simply open a webpage. The webpage has to be served. I found that some versions of some browsers simply ignore this, though, and execute DEIVA even without a web server running.

If that is too difficult or too convoluted, simply go to http://hypercubed.github.io/DEIVA/#/ and drop your file onto the blue button labelled "Select or drop data here". Or just press the button and navigate to the file in question on your hard disk.

Please ask if you have further questions.

Anton