• Module 01: Reproducibile Research
• Module 02: Workshop materials, RStudio Interface, Getting Started with R
• Module 03: Understanding R, Working with objects, R Scripts, R Packages
• — BREAK 10 minutes —
• Module 04: Creating Documents with R Markdown
• Module 05: Create Document = R Script + R Markdown
• Module 06: Customizing R Markdown (templates, parameters and automation)

http://www.nytimes.com/2011/07/08/health/research/08genes.html

2010 Video Presentation by Keith A. Baggerly http://videolectures.net/cancerbioinformatics2010_baggerly_irrh/

• MS Excel inadvertently converts gene symbols to dates and floating-point numbers
• For example, gene symbols such as SEPT2 (Septin 2) and MARCH1 [Membrane-Associated Ring Finger (C3HC4) 1, E3 Ubiquitin Protein Ligase] are converted by default to ‘2-Sep’ and ‘1-Mar’, respectively.
• RIKEN identifiers were automatically converted to floating point numbers (i.e. from accession ‘2310009E13’ to ‘2.31E+13’).
• And gene symbols were converted to dates in supplementary data of recently published papers (e.g. ‘SEPT2’ converted to ‘2006/09/02’).

• In 2014, RStudio released rmarkdown to extend the markdown language originally intended to write documents for the “web” (i.e. HTML).https://daringfireball.net/projects/markdown/syntax

• rmarkdown leverages Pandoc (“universal document converter”) http://pandoc.org/index.html to convert between formats: from HTML (readable by web browsers) to DOC (such as from Microsoft Word or Google Docs) to ODT (Libre Office) to PDF (portable document format) to others like EPUB (e-books), HTML5 slide shows (slidy, ioslides), and TeX based documents and slides (Beamer).

…often called the Swiss-Army knife for converting files from one format to another. Pandoc can convert documents in markdown, reStructuredText, textile, HTML, DocBook, LaTeX, MediaWiki markup, TWiki markup, OPML, Emacs Org-Mode, Txt2Tags, Microsoft Word docx, LibreOffice ODT, EPUB, or Haddock markup to

• Organization
• Clear Documentation
• Standardized
• Centralized
• Efficiency

• What do you want to automate?
• What could you re-use?
  • code, files, formatting, graphics, logos, header, footer, boilerplate
• What should you share with your team?
• What do you find yourself doing over and over?
  • correcting or reformatting
• If you won the lottery today (and left your job), what do you need to tell your replacement so they can pick up where you left off and complete your current tasks?

538.com http://fivethirtyeight.com/ hosts stories and opinion pieces covering poll analyses, politics, economics, health, popular culture, and sports. The founder, Nate Silver, and the 538 team are best known for their political polling and forecasting during the United States Presidential and related elections since 2008. ESPN now owns 538.com (as of 2013) retaining Nate Silver as the Editor-in-Chief.

Most of their articles provide references and links to the original data sources plus details on how their figures, analyses and statistical models were developed. They also host the data, code and details behind their analyses on Github https://github.com/fivethirtyeight/.

We will work with some of these datasets in our exercises later in this course and work with the fivethirtyeight R package https://cran.r-project.org/web/packages/fivethirtyeight/.

Andrew Flowers (economist, data scientist, journalist and former writer for fivethirtyeight.com) presented “Finding and Telling Stories with R” at the 2017 RStudio Conference (Orlando, FL).

The webinar recording of his presentation is available online https://www.rstudio.com/resources/videos/finding-and-telling-stories-with-r/.

In his presentation, he highlights the various aspects of “data journalism” and importance of workflow, data processing and transparency in analysis and communication - all key aspects of reproducibility. Andrew Flowers is also a contributor to the fivethirtyeight R package.

“Our reproducible research policy is for papers in the journal to be kite-marked D if the data on which they are based are freely available, C if the authors’ code is freely available, and R if both data and code are available, and our Associate Editor for Reproducibility is able to use these to reproduce the results in the paper. Data and code are published electronically on the journal’s website as Supplementary Materials.”

https://academic.oup.com/biostatistics/pages/General_Instructions

Example of an article marked R:

• Air pollution and health in Scotland: a multicity study; by Duncan Lee; Claire Ferguson ; and Richard Mitchell; Biostatistics, Volume 10, Issue 3, 1 July 2009, Pages 409–423, https://doi.org/10.1093/biostatistics/kxp010

In 2001 there was an outbreak of E.Coli 0104:H4 that killed 50 people in Europe http://dx.doi.org/10.5524/100001.

Researchers at BGI (formally the Beijing Genomics Institute) worked in collaboration with the Medical Center in Hamburg-Eppendorf to rapidly sequence the genome of the pathogen. Given the severity of the outbreak, the team announced and released the genome via Twitter to the world-wide community of microbial genomicists.

A Github repository was established https://github.com/ehec-outbreak-crowdsourced/BGI-data-analysis/wiki to “crowdsource” analysis and research to find a treatment.

People started contributing their work in under 24 HOURS and within 5 DAYS!! a bacterial agent was proposed to kill the pathogen. This case, highlights the importance of these methods and work practices not only for speed and efficiency but also in rapidly addressing problems and developing solutions that can save lives.

• main component is text
• well written
• good organization and flow
• easily accessible
• understood by team members at all levels
• data + code + text + figures all combined together [e.g. literate programming]
• at end, formatting styles applied via “markup/markdown”

• projects grow
• supporting documentation and files - numerous
• relationships change and can grow more complex
• need file organization and naming schemes
• file names should be:
  • readable by the computer, easy to search, easy to sort (especially by date and author if needed)
  • human readable with logical naming schemes and contain enough info so human knows what is in the file/what the file is for
  • and short enough to be reasonably manageable
• consider user-based access and security (partitioned by “need to know” [users with editing and write permissions versus users with read-only access]

Research Compendium Example https://github.com/ropensci/rrrpkg

• store and share your data and code – even if it is only for your future reference
• sometimes expectation/requirement of funding agency, publisher
• increased visibility, you as source - default subject matter expert
• speed of collaboration - faster advancement of science, knowledge
• good will with community/public

• Cloud-based “File Storage”
  • Dropbox https://www.dropbox.com/
  • Google drive https://www.google.com/drive/
  • Github (better with version control and tracking) https://github.com/
• Data repositories
  • GenBank https://www.ncbi.nlm.nih.gov/genbank/
  • PDB https://www.rcsb.org/pdb/home/home.do
• In addition to Github
  • Bitbucket https://bitbucket.org/
  • Dryad http://datadryad.org/
  • Figshare https://figshare.com/
  • Zenodo https://zenodo.org/

• Yourself
• Your organization - internal reports
• Journals - articles, manuscripts
• Books
• Blogs/Websites
• RSS feeds
• Rpubs https://rpubs.com/
• Gitbook https://www.gitbook.com/
• Bookdown https://bookdown.org/yihui/bookdown/