GitHub is an effective platform for collaborative and reproducible laboratory research

Getting Started

• This repository is an example to demonstrate the organizational principles discussed in Chen et al., 2024. Please refer to the text for more details.
• In each project's repository, open an Issue to discuss ideas, experiments, analyses, or literature.
  • Create Issue templates (e.g. for Literature, Ideas, Experiments, Analyses) to streamline this process.

Repository Organization

• Experiments, Analyses, Grants, Presentations, and Manuscripts go into separate folders.
• Each project contributor creates a sub-folder within the above folders to store their work.
• Do not store data files in the repository. Instead, create symbolic links from where the data is stored (e.g. server or cloud-based storage) so it can be easily accessed in the future.
• The described structure allows all users to be able push to the main branch without introducing merge conflicts, so we do not generally find added benefit in using branches.

Experiments

experiments
    |--USER
        |--ISSUENUMBER_short_desc.md

Analysis

analysis
    |--USER
        |--ANALYSIS_TYPE (eg. riboseq)
            |--ISSUENUMBER_short_desc
                |--data 
                    |--gencode
                        |--gencode.v26.gtf.gz
                    |--fastq
                        |--SRRnnnnnn.fastq.symbolic_link
                |--scripts
                    |--analyze_riboseq.ipynb
                    |--download_from_sra.ipynb
                    |--Snakefile
                |--annotations 
                    |--sample_annotations.csv
                |--tables
                    |--summary_table_1.csv
                    |--summary_table_2.csv
                |--figures
                    |--summary_figure_1.png
                    |--summary_figure_2.png

Manuscripts

manuscripts
    |--USER
        |--manuscript_short_name
            |--maintext
                |--maintext.md
                |--bibliography.yaml
                |--template.tex
                |--pandoc-svg.py
                |--JOURNAL_STYLE.csl
                |--svg
                    |--figure-1.svg
                    |--figure-2.svg
            |--cover_letter
                |--cover_letter.md
                |--template.tex
                |--signature_image.png
                |--institution_logo.png

Grants & Fellowships

grants
    |--USER
        |--YYYY_grant_short_name
            |--research_proposal.md
            |--bibliography.yaml
            |--template.tex
            |--pandoc-svg.py
            |--JOURNAL_STYLE.csl
            |--svg
                |--figure-1.svg
                |--figure-2.svg
            |--project_summary.md
            |--budget_justification.md

Presentations

presentations
    |--USER
        |--YYYYMMDD_presentation_short_desc
            |--presentation.md
            |--STYLE.css
            |--svg
                |--figure-1.svg
                |--figure-2.svg

Software Installation

• Install Docker and VSCode.
• In VSCode, install the Remote Tunnels extension.
• Use Docker to pull the containers necessary for your analyses. Commonly used containers from our registry include:
  • R
  • Python
  • R and Python
  • Pandoc-latex
• To start a container from the command line on your computer, do: docker pull ghcr.io/rasilab/pandoc-latex:1.3.0

docker run -i -d --name pandoc-latex -v $HOME:$HOME ghcr.io/rasilab/pandoc-latex:1.3.0

• You can use the above container as a standalone environment to run any code from the command line. For example, to compile [manuscripts/rasi/paper_template/maintext/maintext.md]manuscripts/kchen/dicodon/maintext/maintext.md, run:

docker exec -w $(pwd) pandoc-latex:1.3.0 pandoc maintext.md --filter=pandoc-svg.py --citeproc --template=template.tex --metadata-file=pandoc-metadata.yaml --pdf-engine=xelatex -o maintext.pdf

• You can also use any of the above containers from within VSCode. For example, we commonly use the R and Python container for interactive analyses (e.g. Jupyter notebooks) in VSCode from our institution's cluster. See instructions here for how to set this up.

Presentations

• Write your presentations as markdown files following the template here.
• Use vscode-reveal extension to open or export the markdown file as a reveal.js presentation.