Classes of existing metadata change upon projection

[ktrns]

Dear all,

Thank you for developing ProjecTILs!

While I am trying out your package, I noticed that classes of the existing metadata are changed after projection.

ref = ProjecTILs::load.reference.map()
sc_projectils = ProjecTILs::make.projection(sc, ref=ref, skip.normalize=TRUE)

Maybe this is something you can fix?

Best wishes and thanks a lot for your work! Katrin

[ktrns]

Hey there,

Me again, once more. Just to let you know... After the above projection command, the following new metadata columns exist but are empty: Sample, Barcode, percent.ribo, percent.mito, Study, functional.cluster, whereby the last one is filled after I called cellstate.predict.

I will work around this, temporarily store the additional Seurat object, extract the relevant filled metadata columns, and add them to my metadata.

Best wishes & thanks Katrin

[ktrns]

Hey there,

Last post for today :-). I noticed you also change rownames of the metadata of my (query) Seurat object to "Q_".

Ideally, as far as I understand it, you wouldn't change my existing metadata table, and just add new columns, because in this case, I wouldn't have to create a second (big) Seurat object.

Best wishes, thanks & have a nice day Katrin

[sjcarmona]

Hi Katrin,

Thank you for your feedback and for reporting this issue! we'll work on it

[mass-a]

Hello Katrin!

these issues with the metadata classes should be solved in the newest release of ProjecTILs.

Best -m