Hardfiltration_Annovar_hg38/tasks/Hardfiltration.wdl

task Hardfiltration {
	
	File vcf_file
	File vcf_index
	String sample
	String docker
	String disk_size
	String cluster_config
	

	command <<<
		set -o pipefail
		set -e
		java -Dsamjdk.use_async_io_read_samtools=false -Dsamjdk.use_async_io_write_samtools=true -Dsamjdk.use_async_io_write_tribble=false -Dsamjdk.compression_level=2 -Xmx32G -jar /gatk/gatk-package-4.1.0.0-local.jar VariantFiltration -V ${vcf_file} -O ${sample}_hc_filtered.vcf --filter-expression "QD<2.0" --filter-name "LowQD" --filter-expression "MQ<40.0" --filter-name "LowMQ" --filter-expression "FS>60.0" --filter-name "FisherStrandBias" --filter-expression "HaplotypeScore>13.0" --filter-name "HighHapScore" --filter-expression "MQRankSum<-12.5" --filter-name "LowMQRankSum" --filter-expression "ReadPosRankSum<-8.0" --filter-name "LowReadPosRankSum" --cluster-size 3 --cluster-window-size 100 --filter-expression "MQ0 >= 4 && ((MQ0/(1.0*DP))>0.1)" --filter-name "HARD_TO_VALIDATE" --filter-expression "DP < 5" --filter-name "LowCoverage" --filter-expression "QUAL < 30.0" --filter-name "VeryLowQual" --filter-expression "QUAL > 30.0 && QUAL < 50.0" --filter-name "LowQual" 
		cat ${sample}_hc_filtered.vcf |grep "#" > ${sample}_hc_filtered.vcf.header.tmp
		cat ${sample}_hc_filtered.vcf |grep PASS > ${sample}_hc_filtered.vcf.tmp
		cat ${sample}_hc_filtered.vcf.header.tmp ${sample}_hc_filtered.vcf.tmp > ${sample}_hc_PASS.vcf
		>>>

		runtime {
		docker:docker
    	cluster:cluster_config
    	systemDisk: "cloud_ssd 40"
    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
	}

	output {
		File raw_vcf = "${sample}_hc_filtered.vcf"
		File pass_vcf = "${sample}_hc_PASS.vcf"
	}
}