satyanarayan-rao / FeatureREDUCE

Repo for running FeatureREDUCE in a virtual environment

Run featureREDUCE

Repo for running FeatureREDUCE in a virtual environment

Please install Anaconda personal addition for linux.

1. Create a virtual environment run_freduce

conda create -n run_freduce python=3.6

2. Activate the virtual environment run_freduce:

conda activate run_freduce

3. Install required packages

sh install_packages.sh

4. Unzip the FeatureREDUCE package

unzip FeatureREDUCE-master.zip

5. Go to FeatureREDUCE-master/freduce/src and run:

javac FeatureReduce.java
cp *.class ../bin

6. Copy envrc FeatureREDUCE-master/freduce/.envrc

cp envrc FeatureREDUCE-master/freduce/.envrc 

Please see the .envrc file already present there and change the path accordingly. Here are the tips:

• Used the command which R to find the path of R and set it as R_HOME

• You will find rJava in $R_HOME/library/rJava : set that as RJAVA_HOME

• You need to put libjvm.so in LD_LIBRARY_PATH. libjvm.so can be found in you virtual library's lib/server directory

• In your CLASSPATH, JRI.jar should be present. This jar file can be found in $RJAVA_HOME/jri/

7. Source the .envrc file

source FeatureREDUCE-master/freduce/.envrc

8. Run the snakemake pipeline to generate results (first run with -np flag to see if dry run looks good, then remove -np flag and run to generate the result)

snakemake -np  --snakefile run_feature_reduce.smk correlation_and_rsq/{max,mef2b,p53}_trial_{01..10}.tsv --configfile configs/config.yaml -j3