Luminarys / C4-Webpage

http://c4.mocklerlab.org/
MIT License
0 stars 0 forks source link

Add in expression queries #6

Closed Luminarys closed 9 years ago

Luminarys commented 9 years ago

Expression query - a multiple gene query, selecting the genes from the expression table (Zmays_Expression), and returning the expression values.

Ideally, this would appear as line plot (maybe a xy plot), but we'll want to have a lot of flexibility: log2 scale, or not (if possible, right on the page with the graph) moving points around on the x axis

Could be difficult: Grouping points by experimental replicates (I'll tell you which goes with which), and producing replicate-group box-plots

Very difficult: dynamically switching between replicate-grouped box-plots and non grouped data.

hdpriest commented 9 years ago

Allow users to designate on the input page if multiple genes are plotted to the same plot, or individual plots.

Also allow users to specify if each gene should be color-coded, or if they wish to see a grey-scale gene expression trend plot (this is helpful for looking at large sets of co-expressed genes, say, N=150, as a mass of co-expressed lines)

If you can, allow users to specify box-and-whiskers plot, or line plots.

Allow users to designate only one (via radio button) of several following normalization methods: Maximum-normalized (each gene individually scaled so that each data point is a fraction of that genes maximum expression - normalizes expression on the interval 0-1)

Mean normalized (each gene scaled such that each data point is divided by the mean expression value for the data series of that gene) (DEFAULT)

Log 2 scaled

hdpriest commented 9 years ago

Start off with just getting the expression data into a table - but you've pretty much got that on lock, so won't take you long.

Luminarys commented 9 years ago

Additional features:

For line graph: plot sample averages, or un-merged data points

take the mean normalized out of box plots, and allow it as an option for all plots

Outlier removal

D3JS Interactiveness - Show popup for an outlier with information.

Luminarys commented 9 years ago

Log2 normalization will not work properly due to the presence of 0s in the data set as well as many values between 0-1. Should it still be utilized?

hdpriest commented 9 years ago

Yes - set Log_2(0)=0, otherwise, Log2 of values between 0-1 giving negative values is fine.

If you can set it so that it re-scales based on user choices, that's best, so if log2 looks wonky for a particular gene, it can be turned off.

Luminarys commented 9 years ago

Everything except interactiveness has been added, so I'm closing the issue.