Genoscores

This is a simple install script for building a standalone version of Genoscores, from source, in a user home directory.

#!/bin/bash
#
#######################################################################################################
#
# Repeatable installation harness for Genoscores, using the dependency details
# documented here: https://genoscores.cphs.mvm.ed.ac.uk/quickstart#Download-and-install-the-package
#
# You MUST supply the genoscores_api.tar.gz file YOURSELF. It is not available to download publicly.
#
# Original author: John.Snowdon@Newcastle.ac.uk
# August 2024
#
#######################################################################################################

# These are the only variables that need to be set per-user:

# This is where genoscore and all the supporting tools (plink, bedtools, etc) will live
MYPATH=$HOME/genoscore

# The path to the genoscores source file, which you must supply yourself...
GENOSCORES_TARBALL=$HOME/genoscores_0.93.0.9000.api.tar.gz

# Set to 1 to enable, or 0 to skip a section
# The first time you run the installer you will want to run all sections...
BUILD_BEDTOOLS=1
BUILD_LIFTOVER=1
BUILD_PLINK=1
BUILD_RLIBS=1
BUILD_GENOSCORES=1

##############################################################################
# NOTHING IS USER-EDITABLE BELOW THIS LINE
##############################################################################

################################################
#
# Set up installation tree
#
mkdir -p $MYPATH
mkdir -p $MYPATH/bin
mkdir -p $MYPATH/zip
mkdir -p $MYPATH/src
if [ -d $MYPATH ]
then
	cd $MYPATH
else
	echo "ERROR! Unable to create $MYPATH"
	exit 1
fi

echo "Standalone Genoscores installer for Rocket HPC"
echo "=============================================="
echo ""
echo "NOTE!!! ALL TOOLS WILL BE INSTALLED UNDER $MYPATH/bin"
echo ""
echo "If you want to change this location, change the variable MYPATH"
echo "at the top of this script."
echo ""
echo "The path you set should be added to your existing PATH environment variable."
echo "For example, at the end of your .bashrc file add:"
echo "PATH="'$PATH'":$MYPATH/bin"
echo "export PATH"
echo ""
echo "After this install has completed, you must then load the R module:"
echo "module load MariaDB/10.6.4-GCC-11.2.0"
echo "module load R/4.3.1-foss-2022b"
echo ""
echo "... and in an interactive R session, register genoscores as follows:"
echo 'library("genoscores")'
echo "initialise.genoscores()"
echo ""
echo "For further documentation on genoscores, see: https://genoscores.cphs.mvm.ed.ac.uk/quickstart"
echo ""
echo "[Press return to continue]"
echo ""
read

#################################################
#
# Download and build bedtools
#
if [ "$BUILD_BEDTOOLS" = "1" ]
then
	echo ""
	echo ""
	echo "################### BEDTOOLS ########################"
	echo ""
	cd zip
	wget -nc https://github.com/arq5x/bedtools2/releases/download/v2.30.0/bedtools-2.30.0.tar.gz
	cd ..

	cd src
	tar xzf ../zip/bedtools-2.30.0.tar.gz
	if [ -d bedtools2 ]
	then
		cd bedtools2
		make -j2 DESTDIR=$MYPATH/bin install
		cd ..
	else
		echo ""
		echo "ERROR! Unable to find bedtools source tree"
		exit 1
	fi
	cd ..

	cd bin
	if [ -d usr ]
	then
		cp -v usr/local/bin/* .
		rm -rf usr
	else
		echo ""
		echo "ERROR! Unable to find bedtools install tree"
		exit 1
	fi
	cd ..
fi

#################################################
#
# Install liftover
#
if [ "$BUILD_LIFTOVER" = "1" ]
then
	echo ""
	echo ""
	echo "################## LIFTOVER ##########################"
	echo ""
	cd bin
	wget -nc http://hgdownload.cse.ucsc.edu/admin/exe/linux.x86_64/liftOver
	if [ -s liftOver ]
	then
		chmod -v u+x liftOver
	else
		echo ""
		echo "ERROR! Unable to find downloaded liftover binary!"
		exit 1
	fi
	cd ..

	cd zip
	wget -nc http://hgdownload.cse.ucsc.edu/goldenPath/hg18/liftOver/hg18ToHg19.over.chain.gz
	if [ -s hg18ToHg19.over.chain.gz ]
	then
		gzip -f -d hg18ToHg19.over.chain.gz
	else
		echo ""
		echo "ERROR! Unable to find downloaded liftover chain file!"
		exit 1
	fi
	if [ -s hg18ToHg19.over.chain ]
	then
		cp -v hg18ToHg19.over.chain ../bin
	fi
	cd ..
fi

###################################################
#
# Install plink
#
if [ "$BUILD_PLINK" = "1" ]
then
	echo ""
	echo ""
	echo "################## PLINK #############################"
	echo ""
	cd zip
	wget -nc https://zzz.bwh.harvard.edu/plink/dist/plink-1.07-x86_64.zip
	cd ..
	if [ -s zip/plink-1.07-x86_64.zip ]
	then
		cd bin
		unzip ../zip/plink-1.07-x86_64.zip
		cp -v plink-1.07-x86_64/plink plink-1.07-x86_64/test.map plink-1.07-x86_64/test.ped .
		rm -rf plink-1.07-x86_64
		cd ..
	else
		echo ""
		echo "ERROR! Unable to find downloaded zip of plink!"
		exit 1
	fi
fi

##############################################
#
# Install R libraries
#
if [ "$BUILD_RLIBS" = "1" ]
then
	echo ""
	echo ""
	echo "################# R LIBRARIES #########################"
	echo ""

	module purge

	# MariaDB is needed for the R::RMariaDB library to build
	module load MariaDB/10.6.4-GCC-11.2.0
	module load R/4.3.1-foss-2022b 
	Rscript -e 'install.packages(c("RMariaDB", "Rcpp", "bigmemory", "corrplot", "data.table","doParallel", "foreach", "ggplot2", "gtools", "httr", "jsonlite", "RColorBrewer", "yaml", "BH", "RcppEigen"), repos = "http://cran.r-project.org")'

	# The current version of 'Matrix' needs R >= 4.4 ... which we don't have on Rocket
	# This is the last version of 'Matrix' which will compile for R <= 4.3 ...
	Rscript -e 'install.packages("https://cran.r-project.org/src/contrib/Archive/Matrix/Matrix_1.6-5.tar.gz", repos=NULL, type="source")'
fi


###############################################
#
# Install genoscores package
#
if [ "$BUILD_GENOSCORES" = "1" ]
then
	echo ""
	echo ""
	echo "################# GENOSCORES ###########################"
	echo ""

	if [ -s $GENOSCORES_TARBALL ]
	then
		# Load the necessary software modules
		module purge
		module load MariaDB/10.6.4-GCC-11.2.0
		module load R/4.3.1-foss-2022b

		# Install the tarball
		R CMD INSTALL $GENOSCORES_TARBALL

		# Test we can import the library
		Rscript -e 'library("genoscores")'
	else
		echo "ERROR! Could not find: $GENOSCORES_TARBALL"
		echo ""
		echo "Please make sure that you have set the correct path and filename of"
		echo "the Genoscores install tarball in the GENOSCORES_TARBALL variable at"
		echo "the top of this script."
		exit 1
	fi
fi

Back to Advanced Software Topics