Warning: NAs introduced by coercion

[moshuangyi2003-spec]

Hello, I encountered an error while conducting analysis using sequencing data. Could you please tell me if there are any requirements for the data format in this software? I didn't receive any error when uploading the rds data, but it gave me an error during the analysis run.
Thank you very much.

[1] "V5 Seurat Object Detected: server.R #853"
[1] "Is null = F, non-orthogonal"
[1] "Is null = F, non-orthogonal"
[1] "We're here 927"
[1] "Is.null = T, non-OrthogonalObject"
Warning: NAs introduced by coercion
Warning: NAs introduced by coercion
Warning: NAs introduced by coercion
Warning: NAs introduced by coercion
Warning: NAs introduced by coercion
Warning: NAs introduced by coercion

[RobertKSuter]

Hi @moshuangyi2003-spec could you please let us know what the error was? Thank you

[moshuangyi2003-spec]

WarningNASintroducedbycoercionWarning:NASintroduced by coercion
Warning:introduced by coercionNASWarning:Error in AddMetaData.
Seurat: No metadata name provided and could not infer it from metadata object
193:stop
192:AddMetaData.Seurat
190: <reactive: seurat_corradded> [C: \users\mac\AppData\Local\R\win-library\4.3\scFocAL\shiny/server.R#1241]188:.func
185: contextFunc

Hello, the error message above is from my system. I used my own custom classification tags for the metadata. There were no errors during the upload and data preprocessing, but an error occurred during the analysis. Does the software have specific requirements for the metadata format?
Thank you very much.

[RobertKSuter]

Hi @moshuangyi2003-spec, this reads as an error when the application is trying to apply labels of sensitive or resistant to whichever small molecule you are using in the in silico perturbation step. Is this when the issue is occuring? Or is this error occuring immediately after completion of the drug-connectivity scoring?

When you say custom classification tags for the metadata, are you referring to using a different format than obj@meta.data?

Thank you!

[moshuangyi2003-spec]

Hello.
Yes. During the running process, the R page kept outputting the warning "Warning: NAs introduced by coercion". After the running was completed, it outputted "Seurat: No metadata name provided and could not infer it from the metadata object".
My labels are my custom labels, such as 'Hep_1', which exist in the mata.data. When using the V4 version of the Seurat object for direct upload, there was no error in this step. Could it be that underlines are not allowed?
Thank you very much.

[RobertKSuter]

Hi @moshuangyi2003-spec - that's possible. Just to be clear, these metadata are the labels you are using to identify control and test populations in the preprocessing tab?

If the app does not crash with the error, and if I understand this is happening after you select a compound and run in silico perturbation? If so, it is possible you have no cells on either of the sensitive or the resistant side based on the cut-offs set in the sliders. You can adjust these cutoffs to make sure you have cells in both groups. If you select a different small molecule, it may resolve that error without needing to adjust the default cut-offs.

For the dashes in the metadata, a quick check would be to make a new metadata column, i.e.

obj$meta2 <- gsub(obj$orig.meta, pattern = "-", replacement = "", fixed = T)

and give it a shot with that. Thank you! & Thanks for your patience