Add slice functions

[noriakis]

Hello, I added the following functions to dplyr_methods.R. I would be grateful if you could review the changes.

• slice_head()
• slice_tail()
• slice_sample()
• slice_min()
• slice_max()

[stemangiola]

Look for "Seurat" in all files, and please replace ;)

[stemangiola]

I believe an empty SCE is completely fine. Not sure this check is needed.

[noriakis]

Thank you very much for your prompt review. I apologize that the part of the message remained the same as tidyseurat. Also, to accept the empty SCE, all the checks for zero length ID in the slice functions are removed.

[stemangiola]

Thanks! Can I please ask to add the 0-length SCE to the uni tests? As we start with our standard SCE and we slice to result to 0 rows, and see what happens (within the unit tests)

[noriakis]

Thank you for the guidance. I have added the tests to expect ncol()=0. All the functions run without error or warnings, except for slice_sample producing the following warning.

> pbmc_small |> slice_sample(n=0)
# A SingleCellExperiment-tibble abstraction: 0 × 17
# Features=230 | Cells=0 | Assays=counts, logcounts
# ℹ 17 variables: .cell <chr>, orig.ident <fct>, nCount_RNA <dbl>, nFeature_RNA <int>, RNA_snn_res.0.8 <fct>,
#   letter.idents <fct>, groups <chr>, RNA_snn_res.1 <fct>, file <chr>, ident <fct>, PC_1 <dbl>, PC_2 <dbl>,
#   PC_3 <dbl>, PC_4 <dbl>, PC_5 <dbl>, tSNE_1 <dbl>, tSNE_2 <dbl>
Warning message:
In max(count_cells$n) : no non-missing arguments to max; returning -Inf

[stemangiola]

Thank you for the guidance. I have added the tests to expect ncol()=0. All the functions run without error or warnings, except for slice_sample producing the following warning.

> pbmc_small |> slice_sample(n=0)
# A SingleCellExperiment-tibble abstraction: 0 × 17
# Features=230 | Cells=0 | Assays=counts, logcounts
# ℹ 17 variables: .cell <chr>, orig.ident <fct>, nCount_RNA <dbl>, nFeature_RNA <int>, RNA_snn_res.0.8 <fct>,
#   letter.idents <fct>, groups <chr>, RNA_snn_res.1 <fct>, file <chr>, ident <fct>, PC_1 <dbl>, PC_2 <dbl>,
#   PC_3 <dbl>, PC_4 <dbl>, PC_5 <dbl>, tSNE_1 <dbl>, tSNE_2 <dbl>
Warning message:
In max(count_cells$n) : no non-missing arguments to max; returning -Inf

Great. If this warning comes from tidySingleCellExperiment we should take care of it.

[noriakis]

Thank you for the review. I have committed the changes to overcome the warning in slice_sample.
I would be grateful if you could review the relevant part.

[stemangiola]

looks good, does the warning appear in tidyseurat as well?

[noriakis]

Thank you for checking. Seems to be yes, although it results in an error. Should I also replace the part for tidyseurat?

> pbmc_small |> slice_sample(n=0)
Error: No cells found
In addition: Warning message:
In max(count_cells$n) : no non-missing arguments to max; returning -Inf

[stemangiola]

@noriakis Yes, please fix the error in tidyseurat and write a unit test for it.

If this is relative to an empty object I thought you had a stop() call already.

[noriakis]

Thank you for your reply. The error is related to an empty object coming from stop() call in subset.Seurat() as you pointed out, so I will just fix the warnings.