study.bokeh_dimensionality_reduction

[olgabot]

Extend/rework the current DecompositionViz to work with Bokeh (https://github.com/bokeh/bokeh). From talking to @bryevdv, a developer of Bokeh, he suggested we wait until the upcoming 0.7.0 release until we start integrating.

Caveats

• Bokeh does not deal with multiple axes objects in a matplotlib figure, so we have to make separate ones

Desirata

• PCA scatterplot, loadings, and violinplots as separate bokeh plots
• Select genes from the loadings plot and ...
  • highlight their vectors
  • Show their violinplots
• In the violinplots, link the samples together so you can hover over a sample and see its expression in all the other genes' violins
  • And in the PCA scatter!

... I think that's plenty

[bryevdv]

Quick notes: 0.7 should be released Friday. In this release it is possible to update bokeh data sources in the noteboook without running a bokeh-server. (Later it will be possible to update any objects, but for now at least data animation/updating based on IPython interactors is possible)

[olgabot]

Thanks for the update! We'll keep you posted on this issue's progress

---

Olga Botvinnik PhD Program in Bioinformatics and Systems Biology Gene Yeo Laboratory http://yeolab.ucsd.edu/yeolab/Home.html | Sanford Consortium for Regenerative Medicine University of California, San Diego www http://olgabotvinnik.com | blog http://blog.olgabotvinnik.com/ | github http://github.com/olgabot | twitter http://twitter.com/olgabot | linkedin http://www.linkedin.com/in/olgabotvinnik

2014-12-02 15:18 GMT-08:00 Bryan Van de Ven notifications@github.com:

Quick notes: 0.7 should be released Friday. In this release it is possible to update bokeh data sources in the noteboook without running a bokeh-server.( Later it will be possible to update any objects, but for now at least data animation is possible)

— Reply to this email directly or view it on GitHub https://github.com/YeoLab/flotilla/issues/204#issuecomment-65324195.