Addition for projections tab to help understand projection quality

[drorbar]

Two things me and Oren thought about:

1. In the projected fold tab - we should have a tick which allows us to see only fitted genes, without it this is a bit hard to use this tab for further decision-making.
2. In the query tab, we have a table called "Gene metadata", which shows the correction factor. It should also show the correlation between this gene in the atlas and in the query, which is what we care about

@orenbenkiki - did we have anything else that we thought about?

[aviezerl]

1. Fitted in which cell type? @orenbenkiki - Is there somewhere a definition of 'global' fitted genes? Where?
2. @orenbenkiki can you point me to the location of this in the anndata?

[orenbenkiki]

It is per cell type but you can OR them

…

On Tue, 12 Mar 2024 at 11:45 Aviezer Lifshitz ***@***.***> wrote: 1. Fitted in which cell type? @orenbenkiki <https://github.com/orenbenkiki> - Is there somewhere a definition of 'global' fitted genes? Where? 2. @orenbenkiki <https://github.com/orenbenkiki> can you point me to the location of this in the anndata? — Reply to this email directly, view it on GitHub <#154 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAQUXR4GMU76KSNU2R7PH3YX3FDLAVCNFSM6AAAAABEJFWAU2VHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DONJXGE2TG> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[orenbenkiki]

There should also be a global fitted mask

…

On Tue, 12 Mar 2024 at 11:46 Oren Ben-Kiki ***@***.***> wrote: It is per cell type but you can OR them On Tue, 12 Mar 2024 at 11:45 Aviezer Lifshitz ***@***.***> wrote: > > 1. Fitted in which cell type? @orenbenkiki > <https://github.com/orenbenkiki> - Is there somewhere a definition of > 'global' fitted genes? Where? > 2. @orenbenkiki <https://github.com/orenbenkiki> can you point me to > the location of this in the anndata? > > — > Reply to this email directly, view it on GitHub > <#154 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAAQUXR4GMU76KSNU2R7PH3YX3FDLAVCNFSM6AAAAABEJFWAU2VHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DONJXGE2TG> > . > You are receiving this because you were mentioned.Message ID: > ***@***.***> >

[aviezerl]

and what about 2? Is there some correlation measure of query and atlas per gene?

[orenbenkiki]

Ugh, good point. I’ll have to compute it and add to the results.

…

On Tue, 12 Mar 2024 at 12:01 Aviezer Lifshitz ***@***.***> wrote: and what about 2? Is there some correlation measure of query and atlas per gene? — Reply to this email directly, view it on GitHub <#154 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAQUXUB4Q2G6BHZALPQQX3YX3G63AVCNFSM6AAAAABEJFWAU2VHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DONJXGM3DA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[orenbenkiki]

I pushed an update that computes the projected_correlation per gene (in addition to per metacell). As for the fitted genes mask, you were right the 1st time - it is only per type - you have to OR them.

…

On Tue, Mar 12, 2024 at 10:15 AM Oren Ben-Kiki ***@***.***> wrote: Ugh, good point. I’ll have to compute it and add to the results. On Tue, 12 Mar 2024 at 12:01 Aviezer Lifshitz ***@***.***> wrote: > and what about 2? Is there some correlation measure of query and atlas > per gene? > > — > Reply to this email directly, view it on GitHub > <#154 (reply in thread)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAAQUXUB4Q2G6BHZALPQQX3YX3G63AVCNFSM6AAAAABEJFWAU2VHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DONJXGM3DA> > . > You are receiving this because you were mentioned.Message ID: > ***@***.***> >