Research project organisation schema - lmmx/devnotes GitHub Wiki

Previous note taking file format specs I've tried to use have gotten abandoned through awkwardness of sticking with them, but that won't happen if the best option is instantiated in code/software compatibility on up (i.e. making the best option the easiest option).

Directories

using ./ as top level dir for a generic research project:

./code/
./data/
./data/results/
./data/results/images/
./docs/
./notes/
./notes/images/
./readings/
./readings/bio
./readings/bio/originals/
./readings/cs/
./readings/cs/originals/

To set up such a template project simply run:

mkdir -p code/ data/results/images notes/images readings/bio/originals readings/cs/originals

Code files

If it's likely going to be helpful to have separate folders, make sub-dir's for e.g. R and py, or by task, etc. (just don't have a mess of miscellaneous files in one code folder).

Data files

"Data" will usually end up being input and output: the sources of files for bioinformatics research for example being things like SRA (Sequence Read Archive), hg19/mm9 (human/mouse genomes). Processing this data through scripts and server jobs will produce

Example data/ sub-folder recommendations:

  • past_job_files/ if running jobs on a HPC cluster make the task easier the next time you come to write a job script
  • fastQ etc. to store sequence data

Note files

  • Notes on meetings etc. (not a specific piece of literature) are named e.g. 06-01 for June 1 (here, all work will be 2016)
  • Notes on specific papers etc. are named e.g. maddox16_stat-genomics.md
    • If there's a second paper with the same author surname and year, append b (maddoxb16_stat-proteomics.md etc.)
    • use - to separate MM-DD in date and words in title (no spaces, more readable in browser address bar)
    • abbreviate words to keep filenames short and informative (this condensing of info is useful)
    • only use capital letters when informative (e.g. in the title gu15_h2az-marks-yeast-AS-promoters.pdf to mean antisense)

This allows programmatic distinction between literature notes and meeting notes. Importantly, literature notes can then be matched to metadata from a spreadsheet (etc.), and notes/PDFs/supplementary PDFs tied together.

  • using editR as Rmd editor will create cache and figure folders (maybe worth putting in .gitignore if versioning notes directories)