In this repository you'll find the code we'll be using during the CNV Machine Learning workshop.
The lesson will be in R, the code is provided as an RMarkdown file. Here you'll find both the Rmd and the compiled HTML file, as well as the simulated CNVs file on which we'll be basing our analysis.
If you use conda or miniconda, you can create an environment with all necessary packages from the provided YAML file. Otherwise the requirements are as follows:
- R vesrion 4.0 or above
- IGV (https://igv.org/)
The folowwing R pakages:
- data.table
- ggplot2
- rmarkdown
We will skip the longer and most computation intensive step or run only on a subset.
To run the full analysis you will also need these additional packages and software:
- CNValidatron (https://github.com/SinomeM/CNValidatron_fl) and it's dependencies
- QCtreeCNV (https://github.com/SinomeM/QCtreeCNV) and it's dependencies
- biomaRt (https://bioconductor.org/packages/release/bioc/html/biomaRt.html)
- tabix
If you plan on running the code yourself before or after the workshop, please note that these instructions assume you are working on a Linux system, Mac should also work.
NB: This lesson assumes basic knowledge of the Linux command line, GNU coreutils, conda, and R.
Clone the required repositories:
# update the path to an actual folder
cd /some/dir/of/your/choice/
# clone this repo
git clone https://github.com/SinomeM/NSHG_CNVs_workshop_2024.git
# move into the lesson repo
cd NSHG_CNVs_workshop_2024
Optional, install the required software using conda and activate the new environment:
# create new env based on the provided YAML
conda env create -f conda_env_rstudio.yml
# activate environment
conda activate rstudio
You can compile the whole script with rstudio or
rmarkdown::render("01_exploratory_analysis.Rmd") and then check out the resulting
HTML.
Alternatively, you can manually execute the R code chunks one after the other.
NB : Some chunks are marked as "# NOT RUN", these are not meant to be manually executed. Some are rmarkdown settings, other are optional steps.
In this repo there are two main folders 01_homework and 02_live_session.
The idea is for you to try running the code in the homework folder and
try to think how you would solve the assignment. Then, in the live session
we will quickly go through the introduction together and move to how we would
solve the issue. There we have a discussion about which analysis method is best depending
on the goal etc.