adverSCarial

[GhislainFievet]

Update the following URL to point to the GitHub repository of the package you wish to submit to Bioconductor

• Repository: https://github.com/GhislainFievet/adverSCarial

Confirm the following by editing each check box to '[x]'

• [ x] I understand that by submitting my package to Bioconductor, the package source and all review commentary are visible to the general public.

• [ x] I have read the Bioconductor Package Submission instructions. My package is consistent with the Bioconductor Package Guidelines.

• [x ] I understand Bioconductor Package Naming Policy and acknowledge Bioconductor may retain use of package name.

• [x ] I understand that a minimum requirement for package acceptance is to pass R CMD check and R CMD BiocCheck with no ERROR or WARNINGS. Passing these checks does not result in automatic acceptance. The package will then undergo a formal review and recommendations for acceptance regarding other Bioconductor standards will be addressed.

• [ x] My package addresses statistical or bioinformatic issues related to the analysis and comprehension of high throughput genomic data.

• [ x] I am committed to the long-term maintenance of my package. This includes monitoring the support site for issues that users may have, subscribing to the bioc-devel mailing list to stay aware of developments in the Bioconductor community, responding promptly to requests for updates from the Core team in response to changes in R or underlying software.

• [ x] I am familiar with the Bioconductor code of conduct and agree to abide by it.

I am familiar with the essential aspects of Bioconductor software management, including:

• [ x] The 'devel' branch for new packages and features.
• [ x] The stable 'release' branch, made available every six months, for bug fixes.
• [ x] Bioconductor version control using Git (optionally via GitHub).

For questions/help about the submission process, including questions about the output of the automatic reports generated by the SPB (Single Package Builder), please use the #package-submission channel of our Community Slack. Follow the link on the home page of the Bioconductor website to sign up.

[bioc-issue-bot]

Hi @GhislainFievet

Thanks for submitting your package. We are taking a quick look at it and you will hear back from us soon.

The DESCRIPTION file for this package is:

Package: adverSCarial
Title: What the Package Does (One Line, Title Case)
Version: 0.99.0
Authors@R: 
    person("Ghislain", "FIEVET", , "ghislain.fievet@gmail.com", role = c("aut", "cre"),
 comment = c(ORCID = "0000-0002-0337-7327"))
Description:
    adverSCarial is an R Package designed for generating and analyzing the vulnerability of scRNA-seq
    classifiers to adversarial attacks. The package is versatile and provides a format for integrating
    any type of classifier. It offers functions for studying and generating two types of attacks,
    min change attack and max change attack. The min change attack involves making a small modification
    to the input to alter the classification. The max change attack involves making a large modification
    to the input without changing its classification.
    The package provides a comprehensive solution for evaluating the robustness of scRNA-seq classifiers
    against adversarial attacks.
License: MIT + file LICENSE
Encoding: UTF-8
Roxygen: list(markdown = TRUE)
RoxygenNote: 7.2.3
biocViews: Software, SingleCell, Transcriptomics, Classification
Suggests: knitr, RUnit, BiocGenerics, TENxPBMCData
Imports: gtools, stringr, randomForest
VignetteBuilder: knitr
Author: Ghislain FIEVET <ghislain.fievet@gmail.com>
Maintainer: Ghislain FIEVET <ghislain.fievet@gmail.com>

[vjcitn]

Please look at the Title field in your DESCRIPTION. Revise and bump the version and push to your repo.

[vjcitn]

Your vignette is called "doc.Rmd"; please use a meaningful filename for your vignette -- eventually there may be multiple such. It includes shorttitle: "Short title for headers" ... do not use the template without reading through. Please run R CMD check on your package before bumping again and verify that there are no ERRORs.

[GhislainFievet]

Hi, thank you for your messages.

I modified what was asked and I ran the R CMD check command with 0 error and 0 warning.

[GhislainFievet]

The tool I submit is a parameter search, and building the vignette can take up to an hour. In my case I had to run the "ulimit -s unlimited" before the check to increase the available memory.

[vjcitn]

Thanks for this comment. We have to be able to check the package in a modest amount of time. THere is a longtests protocol. You should sharply constrain the search in your vignette, or precompute and use a saved run. RIght now I am seeing

--- re-building ‘adverSCarial.Rmd’ using knitr
Error: processing vignette 'adverSCarial.Rmd' failed with diagnostics:
C stack usage  7969824 is too close to the limit
--- failed re-building ‘adverSCarial.Rmd’

SUMMARY: processing the following file failed:
  ‘adverSCarial.Rmd’

Error: Vignette re-building failed.
Execution halted

on a machine that can build all submissions. Please check carefully and follow the contributor guidelines. Thanks

[GhislainFievet]

Thank you for the advices :-) I did constraint the search for a vignette, and precomputed the others. There is no memory limit issue anymore, and the r cmd check command runs in 4 minutes. I did update my github repo.

What is the next step? Do I have too handle the push to git@git.bioconductor.org:packages/.git yet, to trigger the "10 tasks"?

Best regards.

Ghislain

[lshep]

Until you pass precheck, the package is not on git.bioconductor.org yet. Once a reviewer is assigned then yes you would need to push there to trigger new builds. For now just pushing to your github and commenting back here for us to look at it again is sufficient

[GhislainFievet]

Ok, thank you!

[bioc-issue-bot]

A reviewer has been assigned to your package. Learn what to expect during the review process.

IMPORTANT: Please read this documentation for setting up remotes to push to git.bioconductor.org. It is required to push a version bump to git.bioconductor.org to trigger a new build.

Bioconductor utilized your github ssh-keys for git.bioconductor.org access. To manage keys and future access you may want to active your Bioconductor Git Credentials Account

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "ERROR, WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 5c78b7c36611c5e6b49a952e0991de34aeb9e45a

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 687c1b4af9b41e373b2556829e3c0fd5e5601924

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 3a28f0ae5d5f1d2096cee4e8a8c0304228abfad5

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 3887f60ea728f31b9700e7ee222398590cd9c629

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

Congratulations! The package built without errors or warnings on all platforms.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[DarioS]

The package is for perturbation analysis of classification by single cell RNA data.

• The adverScarial.Rmd vignette uses a TENxPBMCData data set and it compiles slowly on my old laptop computer with 4 GB RAM. Please see Memory section of developer's guide. Vignettes should be able to run quickly even on 32-bit hardware (mine is 64-bit).

• List creation is inefficient. For example

modifications <- list()
modifications[[1]] <- list("perc1")
modifications[[2]] <- list("perc99")

This should be done efficiently as modifications <- c("perc1", "perc99"). Please read Pre-allocate and Fill It also shouldn't be a list because all of the elements have the same data type. A character vector is sufficient.

• The code in the vignette overView_analysis.Rmd is not executed. The output is hard-coded. Please make all code chunks in this vignette executable. Same for advRandWalkMinChange.Rmd and adapt_classifier.Rmd.

• Use camelCase parameters and variabes, not snake_case. Please read Variable Names.

• Because there is more than one vignette document and it is not clear which order the vignettes should be read in a README.md file should be part of the repository to guide users how to read the guide. Please see README File for what to include.

• The first vignette which the user should read needs to contain biological motivation at the beginning of it. Why should readers care about min change attack and max change attack? Is it only a theoretical argument or is it modelling some read-world problem about single cell RNA data?

• The package needs interoperbility with existing Bioconductor classes. It currently only accepts base data structures such as matrix and data.frame as input. At a minimum it should accept SingleCellExperiment and DataFrame. Also, there is no input validitity checking by functions.

  @param exprs a matrix or dataframe of numeric RNA expression

• It is unclear what function MClassifier does. The title of documentation is "Example cell type classifier for the pbmc3k dataset" but the parameters are exprs, clusters, target. Why is this function described as working on only one data set? Couldn't it work on any clustered data set?

• In many cases the parameter descriptions do not match their example values. For example, function minChangeOverview has excl_genes = c() and genes = c(). The definitions of these parameters are "a list of genes to exclude from the analysis" and "a list of genes in case you want to limit the analysis on a subset of genes" and Examples section has genes <- c("CD4", "CD8A"). Note that

> class(genes)
[1] "character"

Please don't use list and character vector interchangeably. They are different types in R language.

• All function documentation files lack a Details section. Most of them should have such a section with a paragraph explaining how a function works. For example, MClassifier's documentation doesn't explain the classification algorithm used. Is it k-Nearest Neighbours? Is it Nearest Shrunken Centroids? The reader doesn't know how classification works and how limited it is.

• for loops are not to be used in Bioconductor packages. For example, for (cell_type in unique(clusters)). Please refer to Vectorize of developer's guide. Please convent into vapply loops.

• Don't use 1:something to create sequences. For example, 1:round(1 / step_change_ratio). It should be rewritten as seq_len(step_change_ratio).

• adverscarial.R is has 1197 lines and almost all of the functions in the package are inside of it. Please put one function into one file, unless there are helper function directly related to a main function. Please refer to Organise Functions into Files.

[GhislainFievet]

Thank you for the review :-) A lot of useful advices, I'm on it.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: de76c4e2cc244fe2d80347f3981b7ff53291bc15

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "ERROR". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: c4fabe468e76b2693ce78879e161810d676539f4

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 3b2b3cfeba724d968cade353e6599ac408d2b2d5

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: c8b0ed7e34e5dbe64c42f8726d6eadb2e005cd35

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

Congratulations! The package built without errors or warnings on all platforms.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: c5daf5c69778aa3b57e9ee0865057067c39c7cd1

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 227bc773db223c643ae77e6ea1b7ed77430e1234

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

Congratulations! The package built without errors or warnings on all platforms.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[GhislainFievet]

Hi,

I modified the package according to the review.

A few points:

• I modified the vignettes to increase the number of evaluated chunk, 2 out of 4 vignettes are fully evaluated, and the other are precomputed as advised by vjcitn in the firsts comments of this issue. I wanted to add real world examples with published data like the pbmc3k dataset, and the Bioconductor CHETAH classifier, but I couldn't make it computing time friendly. Vignettes can now be built quite quickly on any computer.
• I have eliminated most of the for loops in the code, but there are still a few that contain side effects.
• Instead of using modifications <- c("perc1", "perc99") as suggested, I have used modifications <- list(c("perc1"), c("perc99")). This enables users to provide more advanced parameters, such as list(c("perc1"), c("full_row_fct", myFct)). I have included detailed instructions on how to construct these parameters in the vignettes and documentation.
• I have added the DataFrame data structure to all functions, but not SingleCellExperiment. This is because classifiers use different parts of SingleCellExperiment objects, such as the raw or normalized data, ENTREZ or gene symbols. To avoid confusion, users should provide the input data matrix directly.

Best regards.

Ghislain

[DarioS]

• Code still widely uses snake_case e.g. tests_grid, gene_ind, modif_ind, exprs_temp, row_results, etc.

• There still remain unnecessary for loops. For example:

  row_results <- c()
  for (gene_ind in seq_len(length(genes)))
  {
   ...
   row_results <- c(row_results, paste(modifications[[modif_ind]], collapse = " "))
  }

  This is ideally suitable for vapply. Also, seq_len(length(genes)) is unnecessary and could simply be seq_along(genes).

  Similarly, cbind at the end of a for loop to incrementally grow a table is not efficient.

  df_result <- data.frame(todel = unique(clusters))
  for (modif_ind in seq_len(length(modifications)))
  {
    ...
    df_result <- cbind(df_result, attacks_length)
  }

  Consider how to use vapply and do.call(cbind, listOfResultingDataFrames) instead.

• I didn't mean that the input should accept DataFrame. DataFrame is not used for gene expression data because the table will not have multiple variable types but they will all be a single numeric type. A few of your return values are data.frame and that would be suited to DataFrame for more compact display. For example #' @return data.frame results of the classification of all the grid combinations. is a return type of plain data.frame.

• The reasoning about not including SingleCellExperiment support is unclear. Which classifiers use raw, unnormalised data for classification? I don't think it makes sense to do that if there are batch effects or different numbers of RNA-seq reads per sample. The documentation is also vague: "#' @param classifier a classifier in the suitable format" what is "the suitable format" precisely? Please provide information in Details section. There's some information in the vignette but it is not clear:

A classifier function has to be formated as follow to be used with adverSCarial:

classifier = function(expr, clusters, target){

   c("cell type", trust_value)
 }

What is trust_value? Please define the output vector clearly.

• The user shouldn't have to manually convert single cell data by executing matrixFromSCE but this function could be used internally at the beginning of user-facing functions if the user specifies a SingleCellExperiment object to them.
• Lots of hidden code chunks at end of first vignette that do nothing important.

[GhislainFievet]

Ok, thank you, I'll correct these points.

Regards.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 66ca8536198f9875ba4450e96bdf487beb00649c

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "ERROR". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 3b4fc2cd640991f9a533b7268596ea2d729a2208

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 15ed90e2e746c662f69f3c524796f5ce32cd8736

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: ad5a48e5e1c9fc16821ce0a17c7b260e4488bbc9

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 1b5579958040240c8ad393271311e995225d2ce7

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

On one or more platforms, the build results were: "WARNINGS". This may mean there is a problem with the package that you need to fix. Or it may mean that there is a problem with the build system itself.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: fc11a933c2bc9a4987b3d70b6a1eaf97649eee47

[bioc-issue-bot]

Dear Package contributor,

This is the automated single package builder at bioconductor.org.

Your package has been built on Linux, Mac, and Windows.

Congratulations! The package built without errors or warnings on all platforms.

Please see the build report for more details. This link will be active for 21 days.

Remember: if you submitted your package after July 7th, 2020, when making changes to your repository push to git@git.bioconductor.org:packages/adverSCarial to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[GhislainFievet]

Hi, here are my corrections.

• ok for camelCase
• I used the do.call(cbind , myList) with lapply
• The first for loop you mentionned modifies two exterior variables: rowResults and exprsTemp, I considered it as side effect, is it ok?
• I removed the DataFrame from the possible input, and made it the output format for minChangeOverView, maxChangeOverview, advGridMinChange and advRandWalkMinChange
• I added explanation on documentation on how to format a classifier, and added information about what is trust_value (which I renamed score)
• I removed the hidden chunks in vignettes.

[GhislainFievet]

About the SingleCellExperiment and the matrixFromSCE: I would like my tool to be versatile. Users should be able to easily adapt any classifier in the appropriate format.

The CHETAH wants as input a SCE, easily created from raw data, eventhough they use normalized data to avoid batch effects. It is easy to adapt the CHETAH classifier to make it take as input a raw data matrix.

input <- SingleCellExperiment(assays = list(counts = input_counts),
                              reducedDims = SimpleList(TSNE = input_tsne))

## Run CHETAH
input <- CHETAHclassifier(input = input, ref_cells = reference)

The scType classifier takes as input a scaled matrix.

es.max = sctype_score(scRNAseqData = pbmc[["RNA"]]@scale.data, scaled = TRUE, 
                      gs = gs_list$gs_positive, gs2 = gs_list$gs_negative) 

In every cases it is easy to provide a data matrix, and make the classification from it.

It is also more convenient for my other functions. Let's say I accept SCE as input for advMinChange:

• I should add a parameter to specify which gene ID I want to use. The pbmc3k from TENxPBMCData comes with ENTREZ id as default, whereas I need gene symbols for the classifier.
• I should add a parameter to specify whether to send to the classifier the SCE object if it is CHETAH, or the scaled data if it is scType. I also need the information to know on which matrix I should make the RNA expression modifications.

It seems a lot easier to handle matrix of data.

I made the matrixFromSCE function for one specific situation, to show as an example in the vignette. But users should make their own functions to get the right matrix from the objects they use, SCE, Seurat, or raw data from the sequencing.