dn070017 / puma Goto Github PK

Please note that these scripts only run on priors that only have 1 type of regulators (miRNAs), as it assumes the regulators in the prior network are all miRNAs (it doesn't use a miRlist.txt file to distinguish between TFs and miRNAs). for the script that runs on both TFs and miRNAs, please check the lab's GitHub.

1. edit the input file names and paths in "puma_config.m"

2. run the puma_run script from the command line using "bash puma_run.sh &". this script will run puma on all samples and store the output in the directory "mat". These files will be used later in the run_lioness script. the output of the puma run in matlab will be stored in the file "puma.biotin2.hpc.uio.no.log". you can check this file to keep track of the run's progress. an email should be sent to your email address when the run is finished

3. run the lioness_run script from the command line using "bash lioness_run.sh x y &". "x" should be the first sample you want to run it from, and "y" the last sample. in the toy dataset "input/expression.txt" there are 5 samples. you can test this by for example running "bash lioness_run.sh 1 3 &" to run lioness only on the first 3 samples. if you want to run lioness on all samples you can use x=1 y=-1. the output of the lioness runs in matlab will be stored in the file "lioness.biotin2.hpc.uio.no.x-y.log" with x and y corresponding to the input number you entered in the command line. you can check this file to keep track of the run's progress. an email should be sent to your email address when the run is finished. the individual lioness runs will be numbered according to the column in the expression data and will be stored in the folder "output"

4. after running, the matlab files are converted to .txt files ("read_and_print_networks.m"). files should be combined ("combine_networks.R") and the degrees should be calculated ("calculate_degree.R"). there are several bash scripts in the folder that run these all together or separately, for example "read_combine_degree.sh". the "calculate_degree.R" file makes use of a file that contains all sample names ("input/sampleorder.txt"). please note that it is also possible to load .mat files directly into R, so this workflow is rather cumbersome and could be optimized

5. after running all of these, you can delete the .mat files in "output" and "mat", the .txt files in "txt", and "net.RData" and "edges.RData" to save space