Cbioportal: Add Mutations profile to Plots tab

[pieterlukasse]

Requirements are specified in RFC 31_c. Basically, it means adding the mutations profile data in a similar way to what we have now for CNA data in the Plots tab. One of the mockup screenshots from the RFC is shown below:

[schultzn]

Good idea. Will this then also expand to box plots where you can show numerical values in relation to mutation status for a given gene, so a two box box-plot, one for "mutated", one for "not mutated"?

[pieterlukasse]

@schultzn yes, the box plot should also be there, but I was thinking of a box plot with one box per mutation status instead of only "mutated" and "not mutated". Do you think this is a good idea? Or do you think the variant classification is not really relevant in this plot?

[schultzn]

It would be nice to have the variant classification as an option - I am sure users will want the ability to combine all mutations.

On Jun 4, 2018, at 5:47 AM, pieterlukasse [e-mail] wrote:

@schultzn https://github.com/schultzn yes, the box plot should also be there, but I was thinking of a box plot with one box per mutation status instead of only "mutated" and "not mutated". Or do you think the variant classification is not really relevant in this plot?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub https://github.com/cBioPortal/cbioportal/issues/4332#issuecomment-394297434, or mute the thread https://github.com/notifications/unsubscribe-auth/ALV34CYhQKpMm7myTW3SCDlperHq9aAhks5t5QIdgaJpZM4UVN9h.

[jjgao]

@pieterlukasse maybe we can split into two attributes: mutation by type and mutant vs wildtype?

[jjgao]

Discussed this with @pieterlukasse.

One option: add the following options (another dropdown below profile name) when a mutation profile is selected:

Mutations by type: missense, inframe, truncating, wild-type, (not sequenced?)
Mutated vs wildtype: mutated, wild-type, (not sequenced?)
Fusions: fused (?), not fused

Also thinking of support altered vs not altered. Maybe we can add another option in the Profile Type drop-down called "Queried alteration"?

@schultzn let's discuss this in person.

[Sjoerd-van-Hagen]

@adamabeshouse how is it going?

[adamabeshouse]

@Sjoerd-van-Hagen noone has discussed this with me :D

[Sjoerd-van-Hagen]

@adamabeshouse good to know... you are working on the plots tab, right?

@jjgao I think @pieterlukasse discussed this with you right? We would to the mutation count and fraction genome altered and then Adam would work on the mutation type?

[jjgao]

@Sjoerd-van-Hagen yes.

@adamabeshouse please work on this after we roll out 1.15.

[adamabeshouse]

ok i added it to 1.16

[jjgao]

Reviving this. @adamabeshouse could you do the followings?

• Data Type dropdown: Clinical Attribute, Mutation, Copy Number, mRNA, Protein, Methylation (hide when not available)
• When Mutation is selected, the next dropdowns are Mutation Data Profile, Mutation Count by, and Gene
• Mutation Data Profile will usually have only one profile which should be automatically selected
• Mutation Count by will have the following items: Mutation Type, Mutated vs Wild-type. (Treat fusion as one type of mutations for now.)
• Gene dropdown will remain the same.

[jjgao]

@Sjoerd-van-Hagen @pieterlukasse Please comment if you would like to make changes to the design above.

[pieterlukasse]

@jjgao thanks, this is good. One small remark I have is only for Mutation Data profile: afaik there is always only one profile, so perhaps this box can be removed in this case.

[jjgao]

@pieterlukasse I just thought it might be good to keep it consistent with other molecular data types, but don't feel strongly either way.

[adamabeshouse]

@pieterlukasse @jjgao lets omit the profile selection, and use that selector for Mutation Count By

[adamabeshouse]

@pieterlukasse @jjgao what order should we put the mutation types in?

[adamabeshouse]

@pieterlukasse @jjgao for choosing mutation type, should we use the same priority algorithm as the oncoprint?

[pieterlukasse]

@adamabeshouse I would say the same algorithm used to color the samples in the plots tab (which hopefully is the same used by oncoprint as well).

[pieterlukasse]

did some tests now on cbioportal.org and it looks great! Thanks @adamabeshouse, @jjgao and @schultzn