scFeatures

[ycao6928]

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

• Repository: https://github.com/SydneyBioX/scFeatures

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 @ycao6928

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: scFeatures
Title: scFeatures: Multi-view representations of single-cell and spatial data for disease outcome prediction
Version: 0.99.1
Authors@R: 
    c(
    person(given = "Yue",family = "Cao", email = "yue.cao@sydney.edu.au",  role=c("aut", "cre")),
    person(given = "Yingxin",family = "Lin", email = "yingxin.lin@sydney.edu.au",  role="aut" ),
    person(given = "Ellis",family = "Patrick", email = "ellis.patrick@sydney.edu.au",  role="aut" ),
    person(given = "Pengyi",family = "Yang", email = "pengyi.yang@sydney.edu.au",  role="aut" ),
    person(given = "Jean Yee Hwa",family = "Yang", email = "jean.yang@sydney.edu.au",  role="aut" ))
Description: scFeatures constructs multi-view representations of single-cell and spatial data.
License: GPL-3
Encoding: UTF-8
LazyData: true
Roxygen: list(markdown = TRUE)
RoxygenNote: 7.2.1
Remotes: 
    sqjin/CellChat
biocViews: CellBasedAssays, SingleCell, Software
Imports:
    ClassifyR,
    DelayedArray,
    DelayedMatrixStats,
    EnsDb.Hsapiens.v79,
    EnsDb.Mmusculus.v79,
    GSVA,
    Seurat,
    ape,
    caret,
    glue,
    dplyr,
    ensembldb,
    gtools,
    msigdbr,
    plyr,
    proxyC,
    reshape2,
    spatstat.core,
    spatstat.geom,
    tidyr,
    AUCell,
    BiocParallel,
    CellChat,
    SpatialExperiment,
    SummarizedExperiment,
    rmarkdown
Suggests: 
    knitr,
    inline
VignetteBuilder: knitr

[vjcitn]

You have

Remotes:
    sqjin/CellChat

All submissions must constrain their dependencies to packages that are in CRAN or Bioconductor. If CellChat is going to Bioconductor, please have it submitted and reviewed before submitting this one.

[vjcitn]

Additionally, I cannot speak for the build system per se, but on two independent linux systems that have succeeded with over 1000 current Bioc packages, I am seeing

tvjc@visually-civil-zebra:~$ R CMD build scFeatures
* checking for file 'scFeatures/DESCRIPTION' ... OK
* preparing 'scFeatures':
* checking DESCRIPTION meta-information ... OK
* installing the package to build vignettes
* creating vignettes ... ERROR
--- re-building ‘Vignette_associationstudy.Rmd’ using rmarkdown
Warning in png(..., res = dpi, units = "in") :
  unable to open connection to X11 display ''
Warning: The vignette title specified in \VignetteIndexEntry{} is different from the title in the YAML metadata. The former is "Vignette_associationstudy", and the latter is "Vignette - association study of features with conditions". If that is intentional, you may set options(rmarkdown.html_vignette.check_title = FALSE) to suppress this check.
--- finished re-building ‘Vignette_associationstudy.Rmd’

--- re-building ‘Vignette_detail.Rmd’ using rmarkdown
make cmd is
  make -f '/home/stvjc/BBS_space/R-4-2-dist/lib/R/etc/Makeconf' -f '/home/stvjc/BBS_space/R-4-2-dist/lib/R/share/make/shlib.mk' SHLIB='file54a0312495fd.so' OBJECTS='file54a0312495fd.o'

make would use
if test  "zfile54a0312495fd.o" != "z"; then \
  echo gcc -shared -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -L/usr/local/lib -o file54a0312495fd.so file54a0312495fd.o /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so  -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -lR; \
  gcc -shared -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -L/usr/local/lib -o file54a0312495fd.so file54a0312495fd.o /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so  -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -lR; \
fi
Quitting from lines 22-36 (Vignette_detail.Rmd) 
Error: processing vignette 'Vignette_detail.Rmd' failed with diagnostics:
/usr/bin/ld: cannot find /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so: No such file or directorycollect2: error: ld returned 1 exit statusmake: *** [/home/stvjc/BBS_space/R-4-2-dist/lib/R/share/make/shlib.mk:10: file54a0312495fd.so] Error 1
--- failed re-building ‘Vignette_detail.Rmd’

--- re-building ‘Vignette_summary.Rmd’ using rmarkdown
make cmd is
  make -f '/home/stvjc/BBS_space/R-4-2-dist/lib/R/etc/Makeconf' -f '/home/stvjc/BBS_space/R-4-2-dist/lib/R/share/make/shlib.mk' SHLIB='file54a05b06e4fc.so' OBJECTS='file54a05b06e4fc.o'

make would use
if test  "zfile54a05b06e4fc.o" != "z"; then \
  echo gcc -shared -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -L/usr/local/lib -o file54a05b06e4fc.so file54a05b06e4fc.o /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so  -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -lR; \
  gcc -shared -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -L/usr/local/lib -o file54a05b06e4fc.so file54a05b06e4fc.o /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so  -L"/home/stvjc/BBS_space/R-4-2-dist/lib/R/lib" -lR; \
fi
Quitting from lines 23-37 (Vignette_summary.Rmd) 
Error: processing vignette 'Vignette_summary.Rmd' failed with diagnostics:
/usr/bin/ld: cannot find /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so: No such file or directorycollect2: error: ld returned 1 exit statusmake: *** [/home/stvjc/BBS_space/R-4-2-dist/lib/R/share/make/shlib.mk:10: file54a05b06e4fc.so] Error 1
--- failed re-building ‘Vignette_summary.Rmd’

SUMMARY: processing the following files failed:
  ‘Vignette_detail.Rmd’ ‘Vignette_summary.Rmd’

Error: Vignette re-building failed.
Execution halted

Are there specific system configuration steps needed to get this package to build?

[vjcitn]

NB this Warning in png(..., res = dpi, units = "in") : unable to open connection to X11 display '' is ameliorated by running with xvfb-run, it is not an issue; the remaining errors noted arise under xvfb-run.

[vjcitn]

library(inline)
openblas.set.num.threads <- cfunction( signature(ipt="integer"),
                                       body = 'openblas_set_num_threads(*ipt);',
                                       otherdefs = c ('extern void openblas_set_num_threads(int);'),
                                       libargs = c ('/usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so'),
                                       language = "C",
                                       convention = ".C"
)
openblas.set.num.threads(1)

must be done in a portable way ... this is from Vignette_details

[vjcitn]

Finally I'd suggest not using "Vignette_" to prefix the name of the vignette files.

[vjcitn]

Use the package name or something more descriptive.

[vjcitn]

Apropos https://github.com/Bioconductor/Contributions/issues/2815#issuecomment-1271492099

please see https://stackoverflow.com/a/45796357/547331 and the associated https://cran.r-project.org/package=RhpcBLASctl -- thanks to Martin Morgan for this observation.

[vjcitn]

Let us know if you are going to work on this. @vjcitn tagging myself to ensure I hear about it.

[ycao6928]

Dear @vjcitn, Thank you so much for looking into the package! Yes, we are actively working on it this week and addressing all the comments.

[nick-robo]

@vjcitn We have made significant changes. Could you have another look?

[nick-robo]

I think that running install.packages(RhpcBLASctl) might fix it. It is listed in the suggests dependencies, so it might not be installed automatically when building.

Another way to solve the dependency issues might be to run devtools::load_all() in the repository directory.

[vjcitn]

Yes, that was my bad, with the BLASctl failure.... but now we have a package of 127MB that is supposed to be no more than 5MB. So you'll have to discuss with @lshep whether this submission should include an ExperimentData package

[vjcitn]

It seems to be passing check cleanly, so we just have to meet the guidelines.

[nick-robo]

We downsampled the data and the package is now under 5MB.

[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". 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/scFeatures 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: 0e77e169f8adaa118afb879eccd533d4e3e2001c

[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/scFeatures 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: 494098d7b6b9d1e31b3ece63b88af3160e4d6b0f

[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/scFeatures to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[nick-robo]

@PeteHaitch Could you please advise how to address the warning? We are not sure.

Warning: replacing previous import 'dplyr::select' by 'ensembldb::select' when loading 'scFeatures'

[PeteHaitch]

You are importing both in your NAMESPACE: https://github.com/SydneyBioX/scFeatures/blob/master/NAMESPACE#L44-L45

I think you will have to choose to import only one and then refer to the other using :: in your package code. E.g., you could choose to have importFrom(dplyr, select) (or the equivalent via roxygen2) and then use ensembldb::select() everywhere in your package code. Alternatively, you could import neither in your NAMESPACE and use the :: versions everywhere in your package code.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 3cdbf3426a09d48bad978ab6060ca608c9f0ec0d

[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/scFeatures 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: b8aef87b57ef47a03ec6a594cccc231be26f7729

[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/scFeatures to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[nick-robo]

@PeteHaitch Sorry for the delay. The package is now building without errors or warnings.

[nick-robo]

Sorry to ping you @PeteHaitch, but today is the deadline for package review for the next Bioconductor release. Is there anything that you would like us to address in the package?

[PeteHaitch]

Hi @nick-robo,

I'm sorry that I have not had time to review scFeatures before the deadline. scFeatures only got an 'OK' on the build system 3 days ago and there were several packages ahead of yours in the reviewing queue.

While scFeatures won't make the next Bioconductor release (3.16), once it's been through the review process and accepted it will be immediately available in the development branch (3.17). I know this may be disappointing, but unfortunately there is only so much time available this close to the release.

Cheers, Pete

[nick-robo]

No worries! We appreciate all the effort that goes into the reviewing process. Thanks for letting us know.

[PeteHaitch]

I'll return to reviewing packages now that the release of Bioconductor 3.16 is done. However, I also have a Masters thesis to review and abstracts for BioC Asia to review, both with deadlines next week, so I'm afraid I won't be able to start my package reviews until those commitments are behind me. Thanks for your patience.

[PeteHaitch]

Hi @ycao6928 and @nick-robo,

Apologies for the delay, but I'm finally back to having time to review BioC packages.

Unfortunately, scFeatures isn't building on BioC 3.17 (devel branch) due to the classifyR dependency being unavailable (see the build report for ClassifyR). I'm also unable to build ClassifyR on my laptop (macOS with Intel).

Would you please:

1. Co-ordinate with @DarioS (ClassifyR maintainer) to get a version of classifyR building in BioC 3.17 (devel branch).
2. Once you have that, do a version bump of scFeatures and push to the BioC git server to trigger a new build as part of the review process.

I'll then be able to do my review proper.

You might also take this opportunity to clarify the nature of the classifyR dependency. It appears that you only use ClassifyR in one of the vignettes, so it could be a suggested dependency (see https://contributions.bioconductor.org/description.html#depends-imports-suggests-enhances for how to do this properly.). There may be other packages that could similarly be moved into suggested packages.

Cheers, Pete

[PeteHaitch]

@ycao6928 and @nick-robo: would you also please clarify who is maintainer of scFeatures and taking responsibility for the review (this should be the same person). Thanks!

[ycao6928]

Hi @PeteHaitch, many thanks for the suggestion! We will move classifyR to suggested dependency. I am the maintainer and taking responsibility of the review.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 6baa2daa7a380e1345866851526cbf61e97438a7

[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/scFeatures 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: 0fa955d8ecc099351ab68efe8fd564988b80c062

[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/scFeatures to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[nick-robo]

@PeteHaitch I am just helping out with getting things working :)

Got a new version that builds! Not sure about the warning, effectiveLibSizes is not mentioned anywhere in the project.

[PeteHaitch]

Got a new version that builds! Not sure about the warning, effectiveLibSizes is not mentioned anywhere in the project.

I believe the warnings regarding effectiveLibrarySizes is coming from recent/in-progress changes to edgeR:

• https://bioconductor.org/packages/devel/bioc/news/edgeR/NEWS
• https://code.bioconductor.org/browse/edgeR/commits/master

We prefer to review a package that is passing without any warnings or errors on the build report, so please try to resolve these first. If you're having trouble resolving it, you might need to post to the bioc-devel mailing list.

[nick-robo]

I might be wrong, but it seems like the problem is upstream.

edgeR is not used anywhere in scFeatures, but it is imported by dependency of a core dependency. It is not declared in scFeatures' namespace, so the warning (declared in NAMESPACE but not found) cannot be related to us.

r$> pak::local_deps_explain("edgeR")
scFeatures -> SpatialExperiment -> DropletUtils -> edgeR

Additionally, I cannot replicate the warning in a local build on a Linux system (all my packages are updated to the latest versions).

[PeteHaitch]

Okay, I'll test on my macbook tomorrow.

[bioc-issue-bot]

Received a valid push on git.bioconductor.org; starting a build for commit id: 80034c0ec37c206d63a08198e22dda351126b805

[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/scFeatures to trigger a new build. A quick tutorial for setting up remotes and pushing to upstream can be found here.

[PeteHaitch]

Hi @ycao6928,

Thank you for submitting scFeatures to Bioconductor.

I think the package needs a little more work, particularly to improve the documentation (man pages and vignettes), before it can be accepted. It also feels a bit awkward for a Bioconductor package to not have first-class support for Bioconductor data structures (SingleCellExperiment and SpatialExperiment). Yes, there's the makeSeurat() converter function but the main functionality requires a Seurat object as input. It would good in future versions to allow a user to pass a SingleCellExperiment/SpatialExperiment object as input (perhaps doing the conversion on the fly, if necessary, or, better yet, directly extracting the relevant data from the SingleCellExperiment/SpatialExperiment object)

In my review below I have separated the issues into Required and Recommended points that I would ask you to address before the package can be accepted. Would you please provide line-by-line comments to my initial review so that I know what changes I'm looking for in my re-review.

Cheers, Pete

Required

• [ ] Most of the code in the vignettes is unevaluated and so is untested by R CMD check. Bioconductor expects that vignettes contain non-trivial evaluated code chunks and static vignettes are generally not acceptable.
  • When I tried running the vignette code line-by-line (like a new user would) there were several examples that returned an error. Such errors would be caught by R CMD check if the vignette code was evaluated. Some examples of where I encountered errors when running the vignette code:

> feature_gene_prop_celltype  <- run_gene_prop_celltype(data_remove_mito, genes = genes_of_interest)
Error: BiocParallel errors
  1 remote errors, element index: 1
  1 unevaluated and other errors
  first remote error:
Error: Under current subsetting parameters, the default assay will be removed. Please adjust subsetting parameters or change default assay.

> feature_CCI <- run_CCI(data , species = "Homo sapiens" )
Error in run_CCI(data, species = "Homo sapiens") : 
  could not find function "run_CCI"

genes_of_interest <- c("TIGIT", "PDCD1")
feature_gene_prop_bulk <- run_gene_prop(data, genes = genes_of_interest )
Error: Under current subsetting parameters, the default assay will be removed. Please adjust subsetting parameters or change default assay.

scfeatures_result <- scFeatures(data, 
             # input selective feature types to generate.      
             feature_types = c("proportion_raw", "pathway_gsva", "L_stats", "gene_mean_celltype",    "gene_prop_aggregated") , 
             # by default assumes data is "scrna" (single-cell RNA-seq), this is now set to "spatial_p" (spatial proteomics).   
             type = "spatial_p" ,
             # by default assumes "Homo sapiens", this is now set to `Mus musculus`. 
             species  = "Mus musculus"  ,   
             # by default uses the 50 hallmark genes,  now set to user specified pathways.   
             geneset = list("pathway_a" = c("SMA" ,  "Vimentin"  ),  "pathway_b" = c("B7H3" ,"FoxP3"   )) ,
             # by default uses top variable genes to generate the celltype specific gene expression feature category,now set to user defined genes. 
             celltype_genes = data.frame(celltype = c("Macrophages"  , "Mono/Neu" ) , marker = c("CD3" , "p53")) ,
             # by default uses top variable genes to generate the overall aggregated gene expression feature category, now set to user defined genes.    
             aggregated_genes =  c("CD3" , "p53") , 
             # When passing as SingleCellExperiment or SpatialExperiment, by default we use the assay stored in "logcount" 
             assay = "norm",
            # By default we look for the sample info in "sample" column and the celltype info in "celltype" column
             sample = "imageID",
             celltype = "cellType",
      #  If users want to construct features from the spatial metrics category, by default we look for the "x_cord" and "y_cord" column 
             spatialCoords = c("x", "y"),
             # by default uses single core
             ncores = 8)
Error: Cannot find 'cellType' in this Seurat object

• [ ] The man pages (documentation) is too terse and imprecise.
  • Each man page should include well-written:
  • Title, Description and Details (see https://r-pkgs.org/man.html#title-description-details)
  • Arguments (see https://r-pkgs.org/man.html#arguments)
  • Return value (see https://r-pkgs.org/man.html#return-value)
  • Examples (see https://r-pkgs.org/man.html#sec-man-examples)
  • To take one example, the documentation of process_data():
  • The 'Description' just repeats the title, which doesn't explain what pre-processing actually is.
  • There's no real description of arguments (what sort of object is data, what sort of normalisation is applied when normalise = TRUE, etc.)
  • The 'Value' section is incorrect (the return value is actually a Seurat object) and it should be explained what the returned object contains.
  • You might take a look at popular Bioconductor packages such as scater or edgeR for inspiration for the documentation (although it needn't be that detailed).
  • Some of the required documentation detail is in the scFeatures_detail vignette, but it properly belongs in the man pages. The vignette should show users how to use the package as a whole and the man pages how to use individual functions.
• [ ] The 'Value' section of many man pages is incorrect. E.g., run_proportion_raw() returns a data.frame but is documented to return 'a matrix of samples x features'. These details matter when it comes to documenting package code.
• [ ] The 'Introduction' of the each vignette, especially the first or main vignette (which I take to be scFeatures_summary) should read like an abstract; see https://contributions.bioconductor.org/docs.html#vignette-introduction
• [ ] Please give the vignettes informative titles and filenames; see https://contributions.bioconductor.org/docs.html#vignettes
• [ ] Please use BiocStyle for vignette formatting.
• [ ] Simply removing ClassifyR from DESCRIPTION is insufficient because the scFeatures_detail vignette still uses it to demonstrate the utility of scFeatures. Now that ClassifyR is passing builds again, please re-add ClassifyR to the DESCRIPTION if using it in the vignettes or elsewhere in scFeatures.
• [ ] RhpcBLASctl should not be needed or used in a package. Why is it needed (and hidden) in the vignette?
• [ ] Why is inline needed in the vignettes?
• [ ] The formatting in the scFeatures_detail vignette is a bit messy because it includes lots of verbose output (progress bars). Please revise.
• [ ] In the scFeatures_detail vignette you are subsetting a vector outside its bounds; please fix:

> unique(data$sample)[1:5]
[1] "Pre_P7"  "Pre_P24" "Pre_P28" "Pre_P33" NA
> unique(data$sample)[1:5]
[1] "Pre_P7_cond_Responder"  "Pre_P24_cond_Responder" "Pre_P28_cond_Responder"
[4] "Pre_P33_cond_Responder" NA

• [ ] In the scFeatures_associationstudy vignette, please use a temporary directory via tempdir(), rather than the current working directory, to demonstrate report generation.
• [ ] That said, the report generation fails when using tempdir() but should work.

> output_folder <- tempdir()
> run_association_study_report(scfeatures_result, output_folder )
Error in abs_path(input) : The file 'output_report.Rmd' does not exist.
In addition: Warning message:
In normalizePath(path, winslash = winslash, mustWork = mustWork) :
  path[1]="output_report.Rmd": No such file or directory

• [ ] The org.Hs.eg.db package is an unspecified dependency used in the report generation demonstrated in scFeatures_associationstudy. Please include it as a dependency in the DESCRIPTION.
• [ ] There are hardcoded paths in the report demonstrated in scFeatures_associationstudy (https://github.com/SydneyBioX/scFeatures/blob/80034c0ec37c206d63a08198e22dda351126b805/inst/extdata/output_report.Rmd#L228). This means it won't work on anyone else's computer.
• [ ] Please add a BugReports field to the DESCRIPTION (usually a link to the Issues page of the GitHub repo).
• [ ] Bioconductor requires documentation of .rds/.Rdata files in inst/extdata in an inst/script/ directory. See data documentation.
  • [ ] inst/extdata is usually used for 'raw' data, so these data might properly belong under data/ rather than inst/extdata/; see https://contributions.bioconductor.org/data.html.
• [ ] Please add a table of contents to each vignette.
• [ ] All man pages should have runnable examples (see https://contributions.bioconductor.org/docs.html#examples)
• [ ] What are the dev and docs folders? Please justify or remove them from the main branch of the git repo (Bioconductor requests that "Any files or directories for other applications (Github Actions, devtools, etc.) should ideally be in a different branch and not submitted to the Bioconductor version of the package." (see https://contributions.bioconductor.org/general.html?q=unnec#undesirable-files)

Recommended

• [ ] It is strongly recommended to add unit tests. I would suggest starting with the individual scFeatures::run_*() functions that underpin the wrapper scFeatures::scFeatures() function.
• [ ] This may be personal preference, but the function outputs feel the wrong way around (with samples as rows and features as columns). This is the opposite of how 'rectangular' data are usually stored in Bioonductor, e.g., SummarizedExperiment, where rows are features (e.g., genes) and samples are columns. A side effect, is that this leads to very 'wide' objects that don't display very nicely when printed (at least that was my experience with the example data). I recommend at least documenting why you choose to return the results in this orientation.
• [ ] makeSeurat() also accepts Seurat objects (https://github.com/SydneyBioX/scFeatures/blob/709f075578bf01b1823dc39fe1d5617472c3f888/R/wrapper_run_scfeatures.R#L225-L237), but this isn't documented. Please document when a user would need this functionality.
• [ ] Please try to cite relevant literature. E.g., in the scFeatures_detail vignette you write, "the L values between the pairs of proteins are calculated using the L function defined in literature" but no reference is given to the relevant literature.
• [ ] In the scFeatures_detail vignette is the advice, "This can be obtained from performing cell type prediction using reference data, for example, using SCTransform from Seurat (see https://satijalab.org/seurat/articles/spatial_vignette.html)". However, to my understanding, SCTransform is a normlization method, not a cell type prediction method, and the link points doesn't point to how to actually do the cell type prediction as best I can tell. Please clarify.
• [ ] Please consider NOTES raised by BiocCheck::BiocCheck(). Code styling notes can be regarded as suggestions, but other points should be followed or reasons given for not following them.

    * NOTE: Update R version dependency from 4.2.0 to 4.3.0.
    * NOTE: Consider adding the maintainer's ORCID iD in 'Authors@R' with
      'comment=c(ORCID="...")'
    * NOTE: Avoid 1:...; use seq_len() or seq_along()
    * NOTE: Use accessors; don't access S4 class slots via '@' in
      examples/vignettes.
    * NOTE: Consider adding runnable examples to man pages that document exported
      objects.
    * NOTE: Consider adding unit tests. We strongly encourage them. See
      https://contributions.bioconductor.org/tests.html

• [ ] Consider adding a top-level README file.
• [ ] Recommend adding a NEWS file (see https://contributions.bioconductor.org/news.html)
• [ ] Recommend adding a inst/CITATION file (see https://contributions.bioconductor.org/citation.html)
• [ ] Consider adding a package-level man page (see https://contributions.bioconductor.org/docs.html#package-level-documentation).
• [ ] It's strongly recommended to avoid direct slot access with @ or slot() of S4 objects. Instead, use accessor functions. That said, my understanding is that Seurat authors have not followed this advice, so it's unavoidable when interacting with Seurat objects, but please ensure you are doing this when interacting with Bioconductor objects.
• [ ] Why does remove_mito() remove more than just mitochondrial genes? I strongly recommend choosing a more precise name for the function or split the functionality up into remove_mito(), remove_ribo(), etc.

[nick-robo]

Thank you for the very detailed feedback! @ycao6928 and I are going to start working on it, and we will let you know when we have implemented the changes.

[ycao6928]

Dear @PeteHaitch, thank you for the very detailed feedback! We have fixed all the required issues and some of the suggested issues. The line by line comments are pasted below:

Required

• [x] Most of the code in the vignettes is unevaluated and so is untested by R CMD check. Bioconductor expects that vignettes contain non-trivial evaluated code chunks and static vignettes are generally not acceptable.
  • When I tried running the vignette code line-by-line (like a new user would) there were several examples that returned an error. Such errors would be caught by R CMD check if the vignette code was evaluated. Some examples of where I encountered errors when running the vignette code:

> feature_gene_prop_celltype  <- run_gene_prop_celltype(data_remove_mito, genes = genes_of_interest)
Error: BiocParallel errors
  1 remote errors, element index: 1
  1 unevaluated and other errors
  first remote error:
Error: Under current subsetting parameters, the default assay will be removed. Please adjust subsetting parameters or change default assay.

• Fixed: The subsetted data included was causing issues. We have now changed the example data that is included in the package.

> feature_CCI <- run_CCI(data , species = \"Homo sapiens\" )
Error in run_CCI(data, species = \"Homo sapiens\") : 
  could not find function \"run_CCI\"

• Fixed: We have now added this function.

genes_of_interest <- c(\"TIGIT\", \"PDCD1\")
feature_gene_prop_bulk <- run_gene_prop(data, genes = genes_of_interest )
Error: Under current subsetting parameters, the default assay will be removed. Please adjust subsetting parameters or change default assay.

scfeatures_result <- scFeatures(data, 
             # input selective feature types to generate.      
             feature_types = c(\"proportion_raw\", \"pathway_gsva\", \"L_stats\", \"gene_mean_celltype\",    \"gene_prop_aggregated\") , 
             # by default assumes data is \"scrna\" (single-cell RNA-seq), this is now set to \"spatial_p\" (spatial proteomics).   
             type = \"spatial_p\" ,
             # by default assumes \"Homo sapiens\", this is now set to `Mus musculus`. 
             species  = \"Mus musculus\"  ,   
             # by default uses the 50 hallmark genes,  now set to user specified pathways.   
             geneset = list(\"pathway_a\" = c(\"SMA\" ,  \"Vimentin\"  ),  \"pathway_b\" = c(\"B7H3\" ,\"FoxP3\"   )) ,
             # by default uses top variable genes to generate the celltype specific gene expression feature category,now set to user defined genes. 
             celltype_genes = data.frame(celltype = c(\"Macrophages\"  , \"Mono/Neu\" ) , marker = c(\"CD3\" , \"p53\")) ,
             # by default uses top variable genes to generate the overall aggregated gene expression feature category, now set to user defined genes.    
             aggregated_genes =  c(\"CD3\" , \"p53\") , 
             # When passing as SingleCellExperiment or SpatialExperiment, by default we use the assay stored in \"logcount\" 
             assay = \"norm\",
            # By default we look for the sample info in \"sample\" column and the celltype info in \"celltype\" column
             sample = \"imageID\",
             celltype = \"cellType\",
      #  If users want to construct features from the spatial metrics category, by default we look for the \"x_cord\" and \"y_cord\" column 
             spatialCoords = c(\"x\", \"y\"),
             # by default uses single core
             ncores = 8)
Error: Cannot find 'cellType' in this Seurat object

• Fixed: Code was not evaluated because of very long runtimes, but this has been changed and broken chunks have been fixed.

• [x] The man pages (documentation) is too terse and imprecise.

  • Each man page should include well-written:

  • Title, Description and Details (see https://r-pkgs.org/man.html#title-description-details)

  • Arguments (see https://r-pkgs.org/man.html#arguments)

  • Return value (see https://r-pkgs.org/man.html#return-value)

  • Examples (see https://r-pkgs.org/man.html#sec-man-examples)

  • To take one example, the documentation of process_data():

  • The 'Description' just repeats the title, which doesn't explain what pre-processing actually is.

  • There's no real description of arguments (what sort of object is data, what sort of normalisation is applied when normalise = TRUE, etc.)

  • The 'Value' section is incorrect (the return value is actually a Seurat object) and it should be explained what the returned object contains.

  • You might take a look at popular Bioconductor packages such as scater or edgeR for inspiration for the documentation (although it needn't be that detailed).

  • Some of the required documentation detail is in the scFeatures_detail vignette, but it properly belongs in the man pages. The vignette should show users how to use the package as a whole and the man pages how to use individual functions.

Fixed: All exported functions now have Title, Description, Arguments, Return value.
Non - exported functions now have description.

• [x] The 'Value' section of many man pages is incorrect. E.g., run_proportion_raw() returns a data.frame but is documented to return 'a matrix of samples x features'. These details matter when it comes to documenting package code. Fixed: Corrected the return value type.

• [x] The 'Introduction' of the each vignette, especially the first or main vignette (which I take to be scFeatures_summary) should read like an abstract; see https://contributions.bioconductor.org/docs.html#vignette-introduction Fixed: Edited the introduction in scFeatures_summary

• [x] Please give the vignettes informative titles and filenames; see https://contributions.bioconductor.org/docs.html#vignettes

  • The titles and names seems sufficiently informative.

• [x] Please use BiocStyle for vignette formatting. Fixed: All vignettes now use BiocStyle.

• [x] Simply removing ClassifyR from DESCRIPTION is insufficient because the scFeatures_detail vignette still uses it to demonstrate the utility of scFeatures. Now that ClassifyR is passing builds again, please re-add ClassifyR to the DESCRIPTION if using it in the vignettes or elsewhere in scFeatures. Fixed: Re-added ClassifyR.

• [x] RhpcBLASctl should not be needed or used in a package. Why is it needed (and hidden) in the vignette? Fixed It was added to fix a bug that is now forgotten, I removed all references.

• [x] Why is inline needed in the vignettes? Fixed: Same as above. It is no longer needed.

• [x] The formatting in the scFeatures_detail vignette is a bit messy because it includes lots of verbose output (progress bars). Please revise. Fixed: Messages and warnings have been removed from the output.

• [x] In the scFeatures_detail vignette you are subsetting a vector outside its bounds; please fix: Fixed This has been fixed by including different data with the package.

• [x] In the scFeatures_associationstudy vignette, please use a temporary directory via tempdir(), rather than the current working directory, to demonstrate report generation.

  • This vignette is not intended to be run as it simulates report generation and should be done with one's own results.

• [x] That said, the report generation fails when using tempdir() but should work.

> output_folder <- tempdir()
> run_association_study_report(scfeatures_result, output_folder )
Error in abs_path(input) : The file 'output_report.Rmd' does not exist.
In addition: Warning message:
In normalizePath(path, winslash = winslash, mustWork = mustWork) :
  path[1]=\"output_report.Rmd\": No such file or directory

Fixed: We have now used tempdir() rather than the current working directory. The no such file bug is also fixed.

• [x] The org.Hs.eg.db package is an unspecified dependency used in the report generation demonstrated in scFeatures_associationstudy. Please include it as a dependency in the DESCRIPTION. Fixed: Added org.Hs.eg.db to suggests.

• [x] There are hardcoded paths in the report demonstrated in scFeatures_associationstudy (https://github.com/SydneyBioX/scFeatures/blob/80034c0ec37c206d63a08198e22dda351126b805/inst/extdata/output_report.Rmd#L228). This means it won't work on anyone else's computer. Fixed: Removed hard-coded path in favour of system.file call.

• [x] Please add a BugReports field to the DESCRIPTION (usually a link to the Issues page of the GitHub repo). Fixed: Added to the description.

• [x] Bioconductor requires documentation of .rds/.Rdata files in inst/extdata in an inst/script/ directory. See data documentation. Fixed: Added documentation for the example_scrnaseq.rds.

  • [x] inst/extdata is usually used for 'raw' data, so these data might properly belong under data/ rather than inst/extdata/; see https://contributions.bioconductor.org/data.html. Fixed: We have tried this, but found the data directory only allows .Rdata file instead of the .RDS format. Once we convert to .Rdata, the file size expanded from under 3MB to 16MB and became too big to include in the package. Therefore we decided to keep the data in the inst/extdata.

• [x] Please add a table of contents to each vignette. Fixed: Added

• [x] All man pages should have runnable examples (see https://contributions.bioconductor.org/docs.html#examples) Fixed: Added examples to all exported functions.

• [x] What are the dev and docs folders? Please justify or remove them from the main branch of the git repo (Bioconductor requests that \"Any files or directories for other applications (Github Actions, devtools, etc.) should ideally be in a different branch and not submitted to the Bioconductor version of the package.\" (see https://contributions.bioconductor.org/general.html?q=unnec#undesirable-files) Fixed: The docs folder is for the pkgdown website. We have now removed the dev folder.

Recommended

• [ ] It is strongly recommended to add unit tests. I would suggest starting with the individual scFeatures::run_*() functions that underpin the wrapper scFeatures::scFeatures() function.

• [x] This may be personal preference, but the function outputs feel the wrong way around (with samples as rows and features as columns). This is the opposite of how 'rectangular' data are usually stored in Bioonductor, e.g., SummarizedExperiment, where rows are features (e.g., genes) and samples are columns. A side effect, is that this leads to very 'wide' objects that don't display very nicely when printed (at least that was my experience with the example data). I recommend at least documenting why you choose to return the results in this orientation. Fixed This is one of the reasons why scFeatures was created.' Rectangular' data are not suitable for statistical modelling, which requires standard sample by feature design matrices. The reason why the output is as such is that the purpose of the features is statistical modelling.

• [x] makeSeurat() also accepts Seurat objects (https://github.com/SydneyBioX/scFeatures/blob/709f075578bf01b1823dc39fe1d5617472c3f888/R/wrapper_run_scfeatures.R#L225-L237), but this isn't documented. Please document when a user would need this functionality. Fixed: Documented the purpose in the the function description, as added examples of cases when user would need this functionality in the vignettes.

• [x] Please try to cite relevant literature. E.g., in the scFeatures_detail vignette you write, \"the L values between the pairs of proteins are calculated using the L function defined in literature\" but no reference is given to the relevant literature. Fixed: Added references for Moran's I and L function.

• [x] In the scFeatures_detail vignette is the advice, \"This can be obtained from performing cell type prediction using reference data, for example, using SCTransform from Seurat (see https://satijalab.org/seurat/articles/spatial_vignette.html)\". However, to my understanding, SCTransform is a normlization method, not a cell type prediction method, and the link points doesn't point to how to actually do the cell type prediction as best I can tell. Please clarify. Fixed: Update the link to https://satijalab.org/seurat/articles/spatial_vignette.html#integration-with-single-cell-data-1, which corresponds to the section using SCTransform for "probabilistic transfer of annotations from a reference to a query set".

• [ ] Please consider NOTES raised by BiocCheck::BiocCheck(). Code styling notes can be regarded as suggestions, but other points should be followed or reasons given for not following them.

    * NOTE: Update R version dependency from 4.2.0 to 4.3.0.
    * NOTE: Consider adding the maintainer's ORCID iD in 'Authors@R' with
      'comment=c(ORCID=\"...\")'
    * NOTE: Avoid 1:...; use seq_len() or seq_along()
    * NOTE: Use accessors; don't access S4 class slots via '@' in
      examples/vignettes.
    * NOTE: Consider adding runnable examples to man pages that document exported
      objects.
    * NOTE: Consider adding unit tests. We strongly encourage them. See
      https://contributions.bioconductor.org/tests.html

• [ ] Consider adding a top-level README file.

• [ ] Recommend adding a NEWS file (see https://contributions.bioconductor.org/news.html)

• [x] Recommend adding a inst/CITATION file (see https://contributions.bioconductor.org/citation.html) Fixed: This is now added

• [ ] Consider adding a package-level man page (see https://contributions.bioconductor.org/docs.html#package-level-documentation).

• [ ] It's strongly recommended to avoid direct slot access with @ or slot() of S4 objects. Instead, use accessor functions. That said, my understanding is that Seurat authors have not followed this advice, so it's unavoidable when interacting with Seurat objects, but please ensure you are doing this when interacting with Bioconductor objects.

• [ ] Why does remove_mito() remove more than just mitochondrial genes? I strongly recommend choosing a more precise name for the function or split the functionality up into remove_mito(), remove_ribo(), etc.

[PeteHaitch]

Thanks for providing your response, @ycao6928. I'm on leave until January 10 and will resume reviews then. Wishing you a good Christmas and New Year break.

[ycao6928]

@PeteHaitch Thank you and happy New Year to you too!