genbankr

[gmbecker]

• 1. What does this package do? (explain in 50 words or less)
     Parses GenBank and GenPept files into a useful datastructure which integrates with the Bioconductor ecosystem.
• 2. Paste the full DESCRIPTION file inside a code block below.

Package: genbankr
Version: 1.1.3
Title: Parsing GenBank files into semantically useful objects
Description: Reads Genbank files.
Authors@R: as.person(c(
      "Gabriel Becker <becker.gabriel@gene.com> [aut, cre]",
      "Michael Lawrence <lawrence.michael@gene.com> [aut]"))
Copyright: Genentech, Inc.
Depends:
    methods
Imports:
    BiocGenerics,
    IRanges,
    GenomicRanges(>= 1.23.24),
    GenomicFeatures,
    Biostrings,
    VariantAnnotation,
    rtracklayer,
    S4Vectors,
    GenomeInfoDb,
    Biobase
Suggests:
    RUnit,
    rentrez,
    knitr
Maintainer: Gabriel Becker <becker.gabriel@gene.com>
VignetteBuilder: knitr
License: Artistic-2.0
RoxygenNote: 5.0.1.9000
biocViews: Infrastructure, DataImport
NeedsCompilation: no

• 3. URL for the package (the development repository, not a stylized html page)
     https://github.com/gmbecker/genbankr
• 4. What data source(s) does it work with (if applicable)?
     NCBI Nucleotide and Protein databases (GenBank and GenPeptfiles)
• 5. Who is the target audience?
     Scientists working on microbial genomics/bioinformatics, on organisms for which NCBI and GenBank are the source of record for annotations.
• 6. Are there other R packages that accomplish the same thing? If so, what is different about yours?
     The rentrez package offers low level querying facilities for all NCBI databases, but does not parse the results into a semantically useful datastructure in the GenBank/GenPept case. I view genbankr as working with rentrez rather than replacing it. In fact, when users give an accession rather than an already downloaded file, rentrez is used to retrieve the raw data, which is then parsed by genbankr machinery.
• 7. Check the box next to each policy below, confirming that you agree. These are mandatory.
  • This package does not violate the Terms of Service of any service it interacts with.
  • The repository has continuous integration with Travis CI and/or another service
  • The package contains a vignette
  • The package contains a reasonably complete README with devtools install instructions
  • The package contains unit tests
  • The package only exports functions to the NAMESPACE that are intended for end users
• 8. Do you agree to follow the rOpenSci packaging guidelines? These aren't mandatory, but we strongly suggest you follow them. If you disagree with anything, please explain.
  • Are there any package dependencies not on CRAN?
  • Do you intend for this package to go on CRAN?
  • Does the package have a CRAN accepted license?
  • Did devtools::check() produce any errors or warnings? If so paste them below.
• 9. Please add explanations below for any exceptions to the above:
     This package is part of the Bioconductor package platform and depends on Bioconductor packages as well as CRAN packages. All dependencies are published in either CRAN or Bioconductor.

This package will not go on CRAN, as it is and will continue to be published as a Bioconductor package (with ROpenSci co-branding, pending success of this submission).

• 10. If this is a resubmission following rejection, please explain the change in circumstances.

[sckott]

Thanks for the submission @gmbecker

[sckott]

Seeking reviewers

[sckott]

Reviewers: @dwinter
Due date: 2016-06-15

[sckott]

@dwinter
Due date: 2016-06-15 - hey there, it's been 21 days, please get your review in soon, thanks 😺 (ropensci-bot)

[dwinter]

Hi @sckott / bot / @gmbecker -- sorry about the delay, I've been working on this but it will take a little longer. Will aim for this week though.

[sckott]

thanks!

[dwinter]

This is a really nice package, which fills an important gap in the tools available to work on sequence data in R. Genbank is a very large sequence database, and genbank-formatted records provide rich information about the sequences it contains. This package provides tools to fetch and parse genbank records, and a set of functions to extract information from the parsed records. Importantly, the paclage integrates well with Bioconductor allowing the information within sequence files to be represented as flexible and useful objects like those provided by GenomicRanges.

Clashes between ropensci and Bioconductor styles

One of the difficulties in reviewing the package is that genbankr is part of the Bioconductor project, and therefore meets different coding standards and practices than those suggested for ropensci. Obviously it's up to the ropensci leaders and @gmbecker to decide how this can be negotiated, I will just list those I find here.

• ropensci suggests using testthat for unit tests, Bioconductor has its own testing framework that uses RUnit.
• ropensci suggest snake_case for most variables, Bioconductor uses camelCase.
• As I understand it, the github repo for this package is a fork of a mirror of the Bioconductor source? I don't know what role the github repo will play for the package and how it can be kept in synch with the "official" source. (I'm sure it can be done, I just wanted to flag it as something that needs to be thought about)

Readme

This is perhaps another example of the different between each project. In ropensci the github repo is a "landing page" for the package, whereas as Bioconductor presents much of the same information here. Even so, the current README for genbankr source code could be improved by adding

• badges for release, continuous integration and test coverage
• Some small usage example
• A clear statement about where to go for help or to file bugs (the Bioconductor QA site?)

I have to say I also find the the installation instructions a little daunting. Perhaps start by emphasizing that almost all users should use the current Bioconductor release, which can be installed easily with biocLite. Can the right version of biocLite be accessed via library(BiocInstaller)? If so, these seems preferable to telling users to source a URL to install packages.

Documentation

I found package-level function to be very good, both clear and complete. Though perhaps not technically documentation, the print function for genbank records is also well thought-out, providing the user wih a clear indication of what each record contains.

The package vignette is also clear and does a good job of demonstrating basic usages and integrations for the package. I think it would also be very helpful to show how the features in a genbank records can be used in an analysis.

For instance, an example demonstrating the best way to use features in a given file and getSeq to retriece only the CDS from an mRNA sequence would document a possible use-case and demonstrate how integration with GenomicRanges can be helpful.

For the example of reading data into and R sessions, I think it's better to store the data in inst/exdata and use system.file to specify the path in vignette. That way a user can copy and paste the example and see it work. (Hadley's book has a section on this)

I appreciate the vignette section of the "limitations" fo the package. The lack of standards as to exactly what goes into a genbank file means such limitations are inevitable, but this clear statement about how those problems are dealt with is helpful.

The code

I have very little to comment on the code itself -- it's well written, clear and appropriately modular. I have not stress tested the package on very large files, but performance on typically-sized files has been good. I don't think there are an missing functions -- the package does a good job of "doing job well", so I like the focus of the current packages is good.

@aahowel3 and I have been using the package in research this week, and we found a couple of records that break the parsing function. These are filed under issue gmbecker/genbankr#1

Everything else

It would be good to have a code of conduct -- even if you envisage mostly working on the project yourself it's useful to have an indication that that would-be collaborators can expect to be treated well.

As mentioned above, it would be useful to have a clear indication of where bugs and/or questions about the package should be sent.

I think the the VignetteBuilder entry in the DESCRIPTION should be changed to rmarkdown, and knitr should be replaced by rmarkdown in the Suggests section too. With these changes I can compile the vignette, run R CMD check and have all tests pass.

Finally, the reviewer recommendations ask for an estimate of the time taken to review the package.
I guess I spend about 2hrs using the package in research (i.e. doing stuff I would otherwise be working on, probably less efficiently 😄), another 2 closely reviewing the code, design and documentation of the package and one hour to prepare this comment.