categorical generic assay feedback

[tmazor]

1. OncoPrint Add Tracks: rename Armlevel CNA to Arm-Level CNA Also in Plots tab:

2. Given that all tracks will have the same values/colors, can there be one legend rather than multiple?

3. Diploid & NA are really hard to tell apart. Can NA be a thin horizontal line like we usually use when data is not available?

4. I tried to add a RNA heatmap but nothing happens when I click 'Add Genes to Heatmap'

5. I added a Microbiome heatmap and it's being plotted like categorical rather than continuous data.

6. I then tried to delete the Microbiome heatmap bu clicking the 3 dots next to the header line, but nothing happens:

7. I assumed NA was for samples without copy number analysis performed (and therefore would be NA for all arms), but this sample is NA for one arm and has a call for a different arm. Is that expected?

8. Change plot axis order to: lost diploid gained

9. Why is NA plotted on the axis? Don't we usually not show samples that are NA for one of the axes?

General questions for @schultzn @jjgao @cBioPortal/product :

10. Given that cancer cells are often not diploid, do we want to use 'diploid' or is it better to use something like 'unchanged' ? And rather than 'gained' and 'lost' should the values be 'Gain' and 'Loss' ? That's slightly more consistent with gene-level CNA terminology.

11. Should we use light blue & pink rather than red and blue?

[dippindots]

@ritikakundra @sbabyanusha Do you know the answer of the comment 7?

7. I assumed NA was for samples without copy number analysis performed (and therefore would be NA for all arms), but this sample is NA for one arm and has a call for a different arm. Is that expected?

[jjgao]

9. Why is NA plotted on the axis? Don't we usually not show samples that are NA for one of the axes?

I suspect that NA was imported and handled as string values. We should handle it the same way when data is unavailable. @dippindots

• make sure the NA is coded in the text file the same way as in other files, e.g. NA in mrna, or gistic_cna
• in the API, NA should be treated the same as other profile data (discarded?)

[tmazor]

Two new issue @dippindots

12. I tried to enter group comparison from a arm-level chart, and the page won't finish loading:
13. If i apply a filter to any other chart (eg select female in sex), i can still see the other slice of the pie (male is still visible, just shaded grey). but if i apply a filter to an arm-level chart, the rest of the chart disappears and i see a full circle of just the slice i selected

Everything listed above seems to be resolved except for (as I was warned): 7, 10, 11

Re 10, 11 - I will summarize here, so that @cBioPortal/product can just read this and hopefully agree.

I think we need to change the wording & coloring used for the arm-level changes. Currently we are using: diploid = Grey gained = Red lost = Blue For gained/lost, we're using the same colors as for Amp/DeepDel -- however, I think that may be misleading and instead we should be using the colors for Gain/ShallowDel

Also since cancer cells are often not diploid, I think we should probably avoid using 'diploid' to indicate the 'unchanged' state.

The screenshot below shows OncoPrint, but the same issues apply to Plots tab and Study View charts.

I propose using the following terms & colors: Unchanged = grey Gain = pink Loss = light blue

[jjgao]

Agreed with all points @tmazor mentioned above.

@dippindots is 13 easy to fix?

[dippindots]

Thanks @jjgao @tmazor! Then I will try to address all comments 10, 11, 12, 13. And at the same time, let's wait for @cBioPortal/curation investigating 7.

[dippindots]

@jjgao I think 12 and 13 should be easy to fix.