# unpack
tar xvfz simulatedData.tar.gz
cd simulatedData

# all tests done with 
# R version 3.1.1 (2014-07-10)

# code: ForwardGenomics/
# directory percentID/ contains the global and local %id data
# directory phylogenies/ contains the phylogenetic trees


# create the jobs for the neutrally evolving elements. They will run very quickly
# the output will be in the dir: FGoutput/
./createJobs_neut.sh FGoutput/ phylogenies/tree_ancestor.nh > jobListNeutral
# create the jobs for the elements under selection (negatives). This will require a cluster run. 
./createJobs_sel.sh FGoutput/ phylogenies/tree_ancestor.nh > jobListSelection

# execute each job in both jobLists


# now merge the CDS and CNE output 
./mergeFGoutput.sh FGoutput/neutral/
# merge the chunks of elements under selection
./mergeFGoutput.sh FGoutput/selection/
# the result is a file called FGout.all.txt in each directory

# get sensitity at 90% precision and plot
# this directory will contain a file listing the data files for each scenario and which methods to compare
mkdir listMethods
getResults.sh FGoutput/ results/ > jobListResults

# execute each job in this jobList

# get the results as a table
# first we need to replace perfect match by perfectmatch, otherwise the R script cannot parse it correctly
find results -name "*dat" | xargs -i sh -c "sed 's/Branch method/Branchmethod/g' {} > {}1"
find results -name "*dat1" | xargs -i sh -c "sed 's/Perfect match/Perfectmatch/g' -i {}"

find results/loss* -name "*dat1" | xargs -i outputSensPrecScenario.perl {} | sort -k1,1g > results.txt

# plot the sensitivity at 90% precision (Fig 3A)
plotPerf.R results/ plot.sensAt90.eps