server-scripts/qhtcp-workflow/docs/REMc_GTF_Ontologies_and_Associated_Heatmaps.txt

This section is derived from the earliest work of Jinyu Guo.  As such it uses Perl scripts. Without porting these two Perl scripts into a new intergrated R script or Python script, one is contrained to use the rather crude copy-paste and and shell script inherient in the original procedures. These two Perl scripts are analyze_v2.pl and terms2tsv_v4.pl which were written in 2003 by Gavin Sherlock for the SGD gene ontology system and require perl installations of such files.  These also require that the gene_ontology_edit.obo, SGD_features.tab files used in the ../Code also be included here. Without rewriting the code, one must compromise directory convenience.  

That being said:
1. One must first execute the Java based WEKA-REMc utility in ones home directory or on ones personal computer and copy the results to the ../REMc/REMcOutput folder. 
2. Next run the DconJG.py script to produce a /GTF_REMcReady folder with all the GTF files extracted from the REMc output files. Be sure to call with python2 (not python).
 
[jwrodger@hartmanlab Code]$ python2 DconJG.py ./REMcOutput/... .csv-finalTable.csv 8 ./

The first argument is the relative path to the ... .csv-finalTalble.csv.  It seems that the script uses the csv file name to create a new directory for the output. The 8 or 10 value is the column number that is needed. The ./ puts that new directory in the working directory. (One could put it in one of the ontolgy directories using a relative path argument.)  

After the script has run, copy that directory with contents just produced to all three ontology directories ../REMc/GTFCode_Ontologies/(Component, Function and Process).

The StudyTemplate directory structure should contain the necessary scripts and files required to run those scrips. Hopefully, on can then go to each ontology subfolder and execute the shell script in a terminal local to the ontology. The Components_... .sh within the /Components folder will call the perl scripts that are local to the /Components folder.  Execute the shell scripts simialarly for /Process and /Function ontologies. All the files are basically the same and just copies. Again without rewriting scrips, the copy paste  redundency is neccessary. 

Next execute the Concatenate_GTF_results.py script from each ontology folder. This will produce an output file of the user's naming within the the ontology folder.

[jwrodger@hartmanlab Process]$ python2 Concatenate_GTF_results.py GTF_REMcReady/ ProcessResults.txt

This script concatenates the from that directory into a single file in this case ProcessResults.txt. but could be any name of the user's choosing. (Again there can be a lack of conformity due to the nature of the old scripts which are used.)