Docker image feasibility

[marencc]

@helixscript @cnobles @chasberry
##Hi Developers!!!
The intSiteCaller and the intSiteUploader are great tools. Many thanks for developing them.
I've had various troubles when trying to install them in my HPC... So I decided to give a try making a Docker image from Dockerfile.
I've retrieved the history from the VM, also the packages that are installed in R (I have added the hg19, which is the one I use)...
I am stuck at the blat installation and I am not sure R will retrieve correctly, I hope it does...

I'd like to share with you the development of the dockerfile in case you can help (PLEASE!) with it!
Dockerfile.txt

[marencc]

To follow up. @colindaven @Gmilton425 @helixscript @cnobles @chasberry

I've sucessfully managed to create a working docker image... It's no the best and it has a lot of room for improvement. Also, I just need the first 2 tools (intSiteCaller and intSiteUploader, I don't know if the subsequent work...)

The way in which I could correctly install all the R libraries that are used was using conda (I know...). But, the version of conda that I used doesn't team up well with dockerfile. The turn around was to setup the environment from the docker file, but to activate it on an interactive session. Thus, proceeding with the setupR.R and setupINSPIIRED.R from inside the container.

I'll post here the dockerfile Final_Dockerfile.txt in case its useful or in case someone can correct the limitations with the conda environment. Please feel free to use it. Also, I added a setupR for the hg19genome:

source("http://bioconductor.org/biocLite.R")
biocLite()
biocLite(c('BSgenome.Hsapiens.UCSC.hg19', 'BSgenome.Mmusculus.UCSC.mm10'), suppressUpdates=TRUE, ask=FALSE)

I am also posting the instructions to proceed with the setupR.R

to build:

docker build -t inspiired1 . >> DockerBuild_v0.log 2>&1

once built, to interact with it:

docker run -it inspiired1 bash

to run inside:

#Activate Inspiired environment
conda activate INSPIIREDv2
R --version #Check the correct version of R is installed
ln -s /bin/tar /bin/gtar
Rscript /root/INSPIIRED/bin/setupR.R
cd ..
Rscript /root/INSPIIRED/bin/setupINSPIIRED.R
Rscript /root/INSPIIRED/bin/SetupR_hg19.R

cd /root/INSPIIRED
export INSPIIRED=$(pwd)
echo $INSPIIRED 

nano components/intSiteCaller/intSiteCallerConfig.yaml 

within that yaml, change:
vectorDataPath : /root/INSPIIRED/inputs/demoDataSet/
I1 : ../intSiteValidation/Data/Undetermined_S0_L001_I1_001.fastq.gz/I1 : /root/INSPIIRED/inputs/demoDataSet/Data/Undetermined_S0_L001_I1_001.fastq.gz
R1 : ../intSiteValidation/Data/Undetermined_S0_L001_R1_001.fastq.gz/R1 : /root/INSPIIRED/inputs/demoDataSet/Data/Undetermined_S0_L001_R1_001.fastq.gz
R2 : ../intSiteValidation/Data/Undetermined_S0_L001_R2_001.fastq.gz/R2 : /root/INSPIIRED/inputs/demoDataSet/Data/Undetermined_S0_L001_R2_001.fastq.gz

nano /root/INSPIIRED/inputs/demoDataSet/INSPIIRED.yml

within that yaml, change, check the correct paths, as they were hardcoded...
parallelize : no
forceQsubPath : No --> yes

#Identify integration sites.
cd inputs/demoDataSet
Rscript $INSPIIRED/components/intSiteCaller/intSiteCaller.R -j demo
Rscript $INSPIIRED/components/intSiteCaller/check_stats.R

[marencc]

Just an update, I've managed to have the conda env initiated
Final_Dockerfile.txt

[colindaven]

Nice. Why not fork the repository and just add the Dockerfile or whatever to your fork ? Link to the fork from this issue then?