shixiangwang/gistic repository overview

⁠Install GISTIC2 by one line code

I have written two Chinese blogs⁠ for telling readers how to install GISTIC 2.0 (a famous software for copy number analysis) step by step. Recently I realize the installation steps can be implemented automatically, so I write this program.

Now, you can directly run GISTIC on hiplot platform, try it at https://hiplot.com.cn/advance/gistic2⁠.

⁠Install from script (Linux Only)

1. Download GISTIC 2.0 from ftp://ftp.broadinstitute.org/pub/GISTIC2.0
2. Download this repository.

git clone https://github.com/ShixiangWang/install_GISTIC
cd install_GISTIC
chmod u+x install_GISTIC2.sh

1. Run script.

This program is a pure bash script and can be run in the following way.

./install_GISTIC2.sh args1 args2

# args1: the path to GISTIC_x_x_xx.tar.gz file
# args2: the install directory, must be absolute path, not relative path

1. Check example script⁠ to see how to run GISTIC.

⁠Install from docker

Two ways:

⁠Just pull from Docker hub https://hub.docker.com/r/shixiangwang/gistic⁠.

docker pull shixiangwang/gistic

⁠Build the image by yourself.

git clone https://github.com/ShixiangWang/install_GISTIC
cd install_GISTIC
sudo docker build -t gistic:latest .

⁠Run docker image

Click example script⁠ to see how to run GISTIC in Docker.

Run the following command to go into Docker interactive terminal.

sudo docker run -it --rm --entrypoint bash shixiangwang/gistic

If you are using singularity, check this example⁠ to see how to run GISTIC.

⁠Note

unset DISPLAY should be added to run script for avoiding X11 related errors.

This program is tested and currently only support GISTIC 2.0.23, any suggestion or pull request is welcome.

⁠Q & A

1. How to handle error related to "overlap".

You can clean the overlap segments byself firstly or use my written R script preprocess.R⁠.

# Usage: Rscript preprocess.R input.seg [minimal_prob, default 0]
#
# Example: Rscript preprocess.R input.seg 5
# Filter segments with less than 5 probes and then clean overlap segments
# by weighted multiplication

⁠Citation

If you want to thank my work, please cite my recent paper and add the link to this GitHub repo.

Antigen presentation and tumor immunogenicity in cancer immunotherapy response prediction, eLife. https://doi.org/10.7554/eLife.49020⁠

⁠LICENSE

MIT © 2019-2020 Shixiang Wang