Yesterday, a friend of mine asked me if there was a good way to organize folders for a bioinformatics project. That is also a good question to me. As for me, when dealing with a relatively light-weight project such as testing a script for GC content counting or performing simple BLAST, I would prefer to use flat folder.
However, when developing a relatively heavy-weight product such as an algorithm, like microsatellite instability determination or an auto-interpretation algorithm for NGS variants under ACMG guidelines, a flat folder structure is not well enough to support a project like this.
Therefore, it is necessary to organize the project folder seriously. The basic idea for organizing the folder is to guarantee the traceability and clarity. So, I strongly recommend the structure below:
.
├── 00.raw
│ └── raw_fasta.fa
├── 01.bin
│ └── process.py
├── 02.ref
│ ├── 20160211_report.pptx
│ └── material.pdf
├── 03.rst
│ ├── publication.png
│ └── svm.xlsx
├── Misc
│ └── MWE
│ ├── demo_input
│ └── work.sh
└── ReadMeThe project folder includes five subfolders, which are raw, bin, ref, rst and Misc. In addition, the ReadMe file should be contained in the project folder. ReadMe file is a form of documentation that contains information about other files in this directory. Each subfolder has its own function in this project.
raw: Save raw files, often come from Illumina machine or websitebin: Save developed scriptsref: Save reference materialsrst: Save important resultsMisc: Save Miscellaneous files. The most important subfolder in it is minimal working example (MWE), which is a collection of source code and other essential data files, allowing a bug or problem to be demonstrated and reproduced.
In a word, a well-organized folder can facilitate human readability and therefore be good for the project. Code is more for computers, and well-organized folder structure is for humans. It is actually often more difficult to deal with humans than computers.