Altered/total copies for somatic, germline SNVs plot; major/minor copies for hetsnps!

[jrafailov]

Data files were delivered to directories - you will see them as mutation_histogram.json. To start, I have ensured that the data will not exceed 1e6 data points. the data has three entries: jabba_cn (an integer), mult_cn (a float), and count (integer).

Please implement equivalently to examples at these links:

https://genome.med.nyu.edu/external/imielinskilab/mskiweb/rafaij01/plots/snvplicity/16323211/altered_copies_histogram_.pdf
You can follow similar paths within https://genome.med.nyu.edu/external/imielinskilab/mskiweb/rafaij01/plots/snvplicity ..

Here are some more specific specs:

1. assign colors to each unique CN present in jabba_cn
2. plot a histogram with the x axis as mult_cn and the y axis as counts
3. plot the histogram such that they are overlayed with one another.. a unique histogram per jabba_cn; not stacked!
4. the plot should render such that upon first looking at the graph all histograms are plotted, but ideally there is a selection mechanism to only view jabba_cn == 1 or 2, for example.

if it is too many data points to begin with, let me know and I can tailor it down. I expect to add more comments and more plots very similar to these within this ticket as followup! thanks in advance!

[shihabdider]

@jrafailov should probably also make an equivalent issue in skilift (assuming we want the code that generates the histogram JSON to be committed there).

[jrafailov]

Hi @xanthopoulakis

some more updates with respect to this issue:

I'm going to ask for you to show 6 individual new plots in the "segment width distribution" tab underneath the total CN purity/ploidy graph. These graphs should be three rows of 2 graphs: somatic, germline, heterozygous SNPs. all data files are now deposited for the nyu clinical samples. I've ensured the # of data points in each of the files will not exceed 1e6 points. I can easily change this parameter if this majorly affects performance especially as we implement this many plots in one view.

All these graphs should be implemented exactly the same way as outlined above.
one note: I think optimal use case should be synced. when an integer CN is selected, that should toggle ALL SIX of the plots changing to only those with jabba_cn == selectedCN shown.

data file references are below and stand for where the plots should sit with respect to another!

SOMATIC: "somatic_altered_hist.json", "somatic_total_hist.json"
GERMLINE: "germline_altered_hist.json", "germline_total_hist.json"
HETSNPS: "hetsnps_major_hist.json", "hetsnps_minor_hist.json"

[jrafailov]

@shihabdider I've created a PR: mskilab-org/skilift#26

[xanthopoulakis]

@jrafailov , I am missing the spec for the bin size in your descriptions. Every histogram needs to specify a bin size for me to calculate and render. The mult_cn attribute is a float, so how do I know what bin size to choose?

[jrafailov]

hey @xanthopoulakis just following up here so its documented:

please add vertical dotted lines at all integer copy states in the six plots you've added.
also please modify "SNVplicity copies" to read "Estimated copies"
lastly the header "SNV Multiplicity" can be edited to just "Multiplicity"
thanks!

[jrafailov]

One more plot to be added!

this one will also go in the segment width distribution and will be called the Purity/ploidy grid search

data is provided for patient id 397089 and is called purple_sunrise.json

we are going to implement analogously to the plot attached within slack

we want the user to be able to interact with the plot. as such, users will "hover" over a particular square on the grid. Each square is represented as a single data point within the json above. Xmin/Xmax define the lower and upperbounds of the ploidy and ymin/ymax define lower and upperbounds of purity.

as always let me know if anything to clarify!