Overview

A quick directory structure and setup for the Methyl Seq Analysis at ARBL.

---

Main files and directories

• R/: contains R code
  • lib.R: contains user defined functions to be used during analysis
  • setup.R: contains package installation instructions (must only be run once)
  • run.R: contains code that runs the code in lib.R once setup.R has been installed fully
• data/: all source data to be used during analysis
  • *.bam: source DNA sequence files stored as *.bam files
  • *.csv: any clinical data that needs to be merged (regarding samples?)
• output/: generated output
  • plots/: contains any plots that may be generated by run.R
  • tables/: contains any *.csv, *.txt, or table-like output from run.R
• reports/:
  • contains analysis output as reports (e.g. results from *.Rmd files)

Example Structure

|--methyl-seq-analysis.Rproj    # RStudio file marks project root
|--R/
    |--setup.R                  # everything necessary to recreate analysis environment
    |--lib.R                    # analysis-specific function library
    |--run.R                    # calls `lib.R`; runs & generates all plots/tables
|--data/
    |--dna-seq-01.bam           # source DNA sequence data
    |--dna-seq-02.bam           # source DNA sequence data
    |--clinical_meta.csv        # source data
    |--training-data.rds        # derived from `*.bam`
|--output/
    |--plots/
        |--figure1_2018-11-09.pdf  # figs resulting from run.R
        |--figure2_2018-11-09.png
    |--tables/
        |--result_table1.csv       # tables resulting from run.R
        |--result_table2.csv
    |--models/
        |--output_model.rds        # any models used
|--reports/
    |--analysis-report.Rmd      # final report in Rmarkdown
                                # refer `source()` `lib.R` in setup chunk
    |--analysis-report.html     # rendered Rmarkdown in `html`
    |--analysis-report.pdf      # rendered Rmarkdown in `pdf`
|--explore/
    |--dead-end-analyses
    |--go-here
|--misc/
    |--reference_paper.pdf      # project-specific free-from wild-card
    |--analysis_plan.docx       # files you don't want polluting the root

General notes

• code is designed to run from the project root (when cloned)
• methyl-seq data goes in the data/ directory
• R code goes in the R/ directory
• Additional reports (e.g. *.Rmd) files go in the reports/ directory
• relative paths can be used when executing from within the project