adding sample interaction from cohort builder

[jaybee84]

-one caveat: projects a lollipopPlot even if it doesnt represent all the samples selected in the cohort.

For eg. if cohort builder contains samples from 3 different studies, and there is gene var data from only study 1, the lollipopPlot plots data from study1 but does not give explicit indication that samples from study 2 and study 3 were not considered.

We may want to add a disclaimer or restrict the plots so that the plots are representative of all samples (eg use == rather than %in%)

[allaway]

I think this behavior is OK. I suspect (just a guess, though) that users would prefer to see what data we have, even if it's not the full cohort.

I agree that we should add a disclaimer to this effect. It could be something like checking all(specimens() %in% x$specimenID) - if this is false, plot some text near or as a part of the plot...we might even be able to do something that allows people to download the cohort table for each plot if they really want to see what is plotted in each.

[allaway]

LGTM!

I added a validation statement to catch an error when no samples are selected. I should add this to my modules too....

validate(need(length(tumor_sample_bc)>0, "No variant data found. Please modify your cohort."))

Feel free to change the wording of the message. The condition might need to be changed in the future to catch other scenarios that I'm not thinking of now.... If it looks good to you, please merge. Thanks!

[jaybee84]

Ah! like the validation message.... will merge...

Also like the idea of checking if all samples are plotted and putting out a disclaimer if that is not the case. Will add that .