gwaybio / pancan_viz Goto Github PK

signatures from individual genes, pathways, and mutation signatures (Alexandrov et al. 2013) would be great to visualize in plots to determine if they are captured by algorithms

Needs to dynamically change with selection of axis

It is getting unwieldy having to manage, load, and reload all the data in the pancan_plotter/data folder. It will be ideal to simply source a single dataframe or object and then the selections in the shiny app will slice appropriately.

Map random ggplot colors to official TCGA pancan cancer-type colors

It would be nice to see the distribution of samples along feature axes - since there are many overlapping points it becomes hard to tell concentration - as suggested by @ajlee21

Drug entries are sloppy, must standardize names

Need to make bins out of years and age variables

Nice work @gwaygenomics!
Is there any plan to include dimensionality reduction via PCA as was done here?

Should be easy to download plots with click of a button. I will need to use downloadhandler()

Current slide bar solution does not work well - especially when there are less than 300 components to view per algorithm. Need to see alternative, dynamic solution for axis selection.

As also mentioned by @rdvelazquez - we will want to visualize mutation burden as a covariate as well.

Would be great to add different types of genomic data profiled by TCGA (e.g. copy number, miRNA, RPPA) as well as different algorithms used to reduce dimensions (e.g. PCA, ICA, NMF, t-SNE) as options to visualize data

Would be great to add 3D and interactive functionality - perhaps three.js would be helpful.

via StackOverflow