Left panel

[AlistairNWard]

This will probably need to be broken up into smaller issues. Any suggestions are just suggestions and we should probably have some UI sessions with the iobio team / other users to optimise things.

@mvelinder Any thoughts on the left panel when you try and use the app?

1. Max Patient Phenotypes probably won't mean much to people. How about Gene:HPO associations?

• DONE Changed as suggested we can rework as needed

3. People will most likely not have any genes of interest. Even if they do, they will probably start out without any, so seeing "n/a / 1" is almost a guarantee and I think it's pretty ugly. Why not have the column as Genes (GoI) and use 0 if there are no defined genes of interest. This way, you will see 1 (0) which looks a lot better imo. Or maybe they should be their own columns - GoI only showing if any are selected. This would help with sorting (4 below)

• DONE Renders goi renders dynamically if there are any split into multiple columns

4. You can use HOM instead of HOM ALT for the inheritance column. The title should be Zygosity - this isn't mode of inheritance. This is pretty standard. Also, you could put HOM in red so we can quickly spot HOM genotypes.

• DONE HOM or HET shows now rather than the longer tag, Title column changed to Zygosity

6. Can we click on column headers to sort? We have to go to the sort button at the top left to choose between sorting on HPO, overlapped genes, genes of interest. I also want to sort on type or zygosity, so the standard way of doing that would just be to click on the column title

• DONE We can sort now by the columns this feature still can probably use polish this is the prototype of it. I think we will want to understand how users expect "multi sort" feedback to function, IMO.

8. kb should be a subscript in the length, e.g. 1.1Kb or 1.1Mb DONE Changed as suggested
9. It would probably useful to include the chromosome in the left panel DONE Changed as suggested
10. Favourite SVs. If a user is stepping through the variants in the left panel, they may find some they find interesting and want to come back to. We could include a favourite mechanism, e.g. a star like in Mosaic. Then you could sort on the star column to bring all favourite SVs to the top Issue Made

9. I don't think people will know what an UNKNOWN ALT is. A ./1 genotype is still a het, so I would recommend leaving this as HET in the zygosity column. If you select the variant, you can then see the actual genotype

• DONE Left as HET if it is ./1, if it is ./. I put unknown we can change that?

10. There are a few concepts around genes and HPO terms that we want to convey. I think the things we want to show are a) total number of genes the SV overlaps; b) genes of interest the SV overlaps; c) number of genes the SV overlaps that are associated with a proband HPO term; d) how many of the proband HPO terms are associated with a gene the SV overlaps. You have a, b, and d covered, but not c. Maybe we should have 4 columns for these counts (which can all be sorted on)
11. Maybe more than we want, but if we have a few more things in the left menu, it will be too large for some use cases. We could consider having an expanded and a condensed view? For example, in Mosaic, we have a variant table that has different views. The expanded view looks like:

and the condensed view:

Maybe we could do something similar? Expanded view could include the coordinate and the condensed view wouldn't. The type and zygosity could be different in the different views. E.g. HET could be and HOM could be , for example. I'm not sure how much space we'd save if the column width is driven by the title, but worth considering

[emersonlebleu]

(7) given own issue
(4) given own issue

[mvelinder]

I'm testing on here, so things may have changed: https://mosaic-staging.chpc.utah.edu/sv.iobio/frontend/

1. Agreed. Max... is confusing. Actually so confusing I'm not questioning what it actually is saying. This is the number of phenotypes that genes are associated with of the total phenotypes loaded? Could you just say Phenotype Associations and then use the syntax X of Y like 4 of 6? The / adds confusion in my opinion.
2. Agreed. GoI was not ever that clear to me. Once I've seen it enough it makes sense. But not intuitive to a new user I'd say. And agreed the n/a is confusing.
   2.5. This "header" of "SV Overlaps" and hierarchy was never very clear to me. Either rework it to be more clear, but I'd suggest thinking more about 9. and what is really the critical overlaps and numbers you want to show and make them more prominent (and sortable descending columns as Al said)
3. Agreed on both
4. That'd also be nice
5. Yep
6. Agree
7. See below
8. Agreed, had no idea what that genotype/zygosity was
9. Agree, there are some critical numbers we want to know, but right now they're all a bit confused (see 1 and 2). A rework of this to be really clear about the important associations would be a big improvement.
10. More color generally would be nice. Or badges like Mosaic uses, even if they are mostly text based they have a background that provides some color. Maybe saving the horizontal space (below) would help too

• Stars indicate favorites in my mind, right now star is denoting the selected variant. Maybe use a border in the left panel around the entry/row when it's selected? Here and elsewhere I'd suggest looking to gene.iobio for style choices like this. It will help keep the apps consistent in feel too
  
• Not sure where SV Summary column is getting pulled from, but it has redundant information in the VCF I loaded at least. Both TYPE and SIZE are shown twice in the panel. You can save quite a bit of horizontal space by removing these duplicates.
  

[AlistairNWard]

That's a good point on the star. I also agree on consistency. We have a family of tools in Mosaic and iobio apps and where we can have consistency between them, we should

[mvelinder]

I'm also not sure what the gene name under the 2/2 here is indicating. That RBM8A is associated with both loaded phenotypes? But in the bottom scroll bar it looks like RPS27 and UBE2T are also both associated with both terms. Is that true? If so, you should indicate there are more than just that one gene. Or maybe not even show the gene name there. We could try to come up with a better way to show that gene name information and more completely

[AlistairNWard]

Do you want to start a new issue? This issue was just about the left panel. I'm going to add additional issues for the other parts of the app

[emersonlebleu]

5. kb should be a subscript in the length, e.g. 1.1Kb or 1.1Mb

@AlistairNWard I'm sorry I'm not sure what you mean here. Is it the lowercase 'k'? It should be capitalized to mirror Mb. But are you not wanting a subscript style on the scaling?

[AlistairNWard]

Sorry - should NOT be a subscript. Lowercase k and upper case M are standard

[emersonlebleu]

• Not sure where SV Summary column is getting pulled from, but it has redundant information in the VCF I loaded at least. Both TYPE and SIZE are shown twice in the panel. You can save quite a bit of horizontal space by removing these duplicates.

@mvelinder Totally right, I just added the summary recently and it does make those other columns redundant in my opinion. I think we can free up that space maybe and add a few of the other more important SV related metrics. Freeing up two columns might allow room for Al's num 9.

[mvelinder]

Do you want to start a new issue? This issue was just about the left panel. I'm going to add additional issues for the other parts of the app

@emersonlebleu this is now #126