12_GATK.srun

#!/bin/bash
#SBATCH -A phillipslab
#SBATCH --partition=phillips       ### Partition
#SBATCH --job-name=GATK    ### Job Name
#SBATCH --time=240:00:00        ### WallTime
#SBATCH --nodes=1              ### Number of Nodes
#SBATCH --ntasks=1             ### Number of tasks per array job
#SBATCH --array=0-19          ### Array index
#SBATCH --cpus-per-task=1            ### Number of CPU cores per task
export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK


module load java easybuild GATK

ref_CE="/projects/phillipslab/ateterina/CITP_pi/ref/c_elegans.PRJNA13758.WS280.genomic.fa"
ref_CB="/projects/phillipslab/ateterina/CITP_pi/ref/c_briggsae.PRJNA10731.WS280.genomic.fa"
ref_CT="/projects/phillipslab/ateterina/CITP_pi/ref/c_tropicalis.PRJNA53597.WS280.genomic.fa"

GATK="/projects/phillipslab/ateterina/scripts/gatk-4.1.4.1/gatk"
cd /projects/phillipslab/ateterina/CITP_pi/data/


LISTFILES=(*f[ps]1.fastq)
file0=${LISTFILES[$SLURM_ARRAY_TASK_ID]}
file=${file0/_f[ps]1.fastq/.ded.bam}


if [[ "$file" =~ "CE".* ]]; then

	echo "C. elegans"
	$GATK --java-options "-Xmx5g -Xms5g" HaplotypeCaller -R $ref_CE -I ${file}  -O ${file/.ded.bam/.raw.g.vcf} -ERC GVCF;

elif [[ "$file" =~ "CB".* ]]; then

	echo "C. briggsae"
	    $GATK --java-options "-Xmx5g -Xms5g" HaplotypeCaller -R $ref_CB -I ${file}  -O ${file/.ded.bam/.raw.g.vcf} -ERC GVCF;

else

	    echo "C. tropicalis"
		$GATK --java-options "-Xmx5g -Xms5g" HaplotypeCaller -R $ref_CT -I ${file}  -O ${file/.ded.bam/.raw.g.vcf} -ERC GVCF;


fi