This repository is meant to serve as a template for bulk RNAseq projects. The goal is to align code across projects and analysts to speed up analyses and ensure consistent, quality science.
Before using this repository, please see the following entries in the knowledgebase for how to set up a successful project smoothly:
Dockerized Rstudio: Isolated containers with consistent architecture, R versions, and machine dependencies for essential libraries (e.g. DESeq2, Seurat).
Renv: Track versions of R libraries used for a project and enable a cache of different versions for different projects.
Format files: Workflow for rendering HTML reports.
Git and GitHub with RStudio (server): Getting git credentials setup in RStudio server, troubleshooting common git/GitHub/RStudio problems, and creating new gencore analysis projects from templates.
Template repos: Directory structure and scripts generalizable to an analysis. These serve as starting points for new analyses.
Gencore packages: Standardized code for common gencore analyses.
Click the green button in the upper-right corner to make a new repo using this as the template, or use GitHub CLI to make a new repo from this template.
- Edit the config files for your study
- Create an experimental design table. The script
helper_scripts/generate_experimental_design_sheet.Rcan be a good starting point. You will need to fill in study metadata. - Use the
analysis_scripts/bulk_rnaseq_qc_template.runfile.Rmdto perform QC and create a standalone object. This script is relatively 'push button' if the config sheets are set up correctly, so you should just be able to run all chunks. The QC template relies on functions from the gencore-bulk repo. See the link for install instructions - Run the model validation scripts to generate an internal report for reviewing model design choices (still a WIP).
- Run the
analysis_scripts/bulk_rnaseq_analysis_template.runfile.Rmd. You will need to set up an experimental formula, and likely make other edits.
A yaml-style config specifies variables to be used in the script. Here you specify the STAR output directory, STAR reference directory (used to rename genes from IDs to symbols), PCA mapping aesthetics, etc.
The template relies on metadata provided in a table format. The table should include the following fields:
-
FileID: sample-specific prefix to the STAR counts file. The redundant part of the file name (e.g. " _ReadsPerGene.out.tab") can be specified in the config as STARreadSuffix
-
SampleID unique identifier for each sample, ideally short and easy to read as it is used for figure labels
Other fields can be supplied as relevant. You can refer to these columns by their name in the QC_config.yml to add layers to plots.
An example design table
| FileID | SampleID | Group | Intgroup | Label |
|---|---|---|---|---|
| p22074-s007_LiproxposFAC1_S147 | p22074-s007 | LiproxposFAC | LiproxposFAC | LiproxposFAC_S147 |
| p22074-s008_LiproxposFAC2_S148 | p22074-s008 | LiproxposFAC | LiproxposFAC | LiproxposFAC_S148 |
| p22074-s012_MTposFAC3_S152 | p22074-s012 | MTposFAC | MTposFAC | MTposFAC_S152 |
| … | … | … | … | … |
| #Required, unique and specific based on file names | #Required, unique, anything | Not required | Not required | Not required |