LIMIX command line tool

Command tool for fitting linear models using csv input dataframes.

• GitHub repo: https://github.com/letaylor/limix_cl
• Free software: MIT license

Features

• Easy to scale jobs over many y features using the fold and iteration options.
• Supports permutation calculations, by permuting the sample ids of the x dataframe.

Quickstart

Install:

pip install git+https://github.com/letaylor/limix_cl

When installing the limix package, you may need to manually install liknorm.

The limix_cl installation automatically includes an execution script which can be run as shown below:

# get help
limix_cl --help

# get the location of test files
dat_dir=`pip show limix_cl | grep Location | awk '{ print $2}'`
dat_dir=${dat_dir}"/limix_cl/tests/data"

# basic call
limix_cl --y_file ${dat_dir}"/y_dat.csv" --x_file ${dat_dir}"/x_dat.csv"

# add covariates and also select a subset of samples
limix_cl --y_file ${dat_dir}"/y_dat.csv" --x_file ${dat_dir}"/x_dat.csv" --covariates_file ${dat_dir}"/cov_dat.csv" --sample_file ${dat_dir}"/sample_subset.txt"

# 100 permutations with seed as 6 (for reproducibility)
limix_cl --y_file ${dat_dir}"/y_dat.csv" --x_file ${dat_dir}"/x_dat.csv" --covariates_file ${dat_dir}"/cov_dat.csv" --permute 100 --permute_seed 6

# split the y file in to 3 batches, with current execution the 1st batch
limix_cl --y_file ${dat_dir}"/y_dat.csv" --x_file ${dat_dir}"/x_dat.csv" --fold 3 --iteration 0

# now run the 2nd batch
limix_cl --y_file ${dat_dir}"/y_dat.csv" --x_file ${dat_dir}"/x_dat.csv" --fold 3 --iteration 1

# now run the 3rd batch
limix_cl --y_file ${dat_dir}"/y_dat.csv" --x_file ${dat_dir}"/x_dat.csv" --fold 3 --iteration 2

# now run the 4rd batch - raises error since there is no 4th batch
limix_cl --y_file ${dat_dir}"/y_dat.csv" --x_file ${dat_dir}"/x_dat.csv" --fold 3 --iteration 4

# list files generated
ls swarm-*

# delete all files generated
rm -r swarm-*

Install Notes

You may find errors relating to machine_ffi or liknorm. This is due to the liknorm package required by limix. You may need to manually install the liknorm headers to locally. Below are commands that should do that (drived from https://liknorm.readthedocs.io/en/stable):

# install liknorm headers locally
git clone https://github.com/glimix/liknorm.git
cd liknorm
mkdir build
cd build
cmake .. -DCMAKE_INSTALL_PREFIX:PATH=$HOME
make
make install

# install liknorm pip, passing the local install of the headers
pip install liknorm --global-option=build_ext --global-option="-I$HOME/include/" --global-option="-L$HOME/lib"

# note: you may also need to update your LD_LIBRARY_PATH
#export LD_LIBRARY_PATH=$HOME/lib:$LD_LIBRARY_PATH