run_processing.sh 13.44 KiB
#!/bin/bash
#PBS -l walltime=24:00:00
#PBS -l ncpus=1,mem=16gb
#PBS -q uv2000
#PBS -N run_processing
#PBS -j oe
# setup PATH
export PATH=$PATH:/home/u027/project/software/bcbio/anaconda/envs/python2/bin:/home/u027/project/software/bcbio/anaconda/bin
export PERL5LIB=$PERL5LIB:/home/u027/project/software/bcbio/anaconda/lib/site_perl/5.26.2
### folder structure for the downstream analysis - created by processing_setup.sh ###
BASE=/scratch/u027/project/analysis/wes_pilot
WORK_DIR=$BASE/${PROJECT_ID}
VCF_DIR=${WORK_DIR}/VCF
PED_DIR=${WORK_DIR}/PED
LOG_DIR=${WORK_DIR}/LOG
G2P_DIR=${WORK_DIR}/G2P
VASE_DIR=${WORK_DIR}/VASE
COV_DIR=${WORK_DIR}/COV
DEC_DIR=${WORK_DIR}/DECIPHER
IGV_DIR=${DEC_DIR}/IGV
CNV_DIR=${WORK_DIR}/CNV
SCRIPTS_DIR=/home/u027/project/scripts
# other files to be used
FAMILY_IDS=${WORK_DIR}/FAM_IDs.txt # created by processing_setup.sh
CHILD_IDS=${WORK_DIR}/PRO_IDs.txt # created by processing_setup.sh
TARGETS=/home/u027/project/resources/DDG2P.20190919.plus15bp.merged.bed # OK
CLINVAR=/home/u027/project/resources/DDG2P.20190919.clinvar.20190916.plus15bp.txt # OK
BLACKLIST=/home/u027/project/resources/current_blacklist.txt # OK
TRANS_MAP=/home/u027/project/resources/current_trans_map.txt # OK
### TOOLS ###
BCFTOOLS=/home/u027/project/software/bcbio/anaconda/envs/python2/bin/bcftools
BGZIP=/home/u027/project/software/bcbio/anaconda/envs/python2/bin/bgzip
TABIX=/home/u027/project/software/bcbio/anaconda/envs/python2/bin/tabix
VT=/home/u027/project/software/bcbio/anaconda/bin/vt
VASE=/home/u027/project/software/bcbio/anaconda/bin/vase
GATK4=/home/u027/project/software/bcbio/anaconda/bin/gatk
GATK3=/home/u027/project/software/GenomeAnalysisTK-3.8/GenomeAnalysisTK.jar
PYTHON2=/home/u027/project/software/bcbio/anaconda/envs/python2/bin/python2.7
VEP="/home/u027/project/software/bcbio/anaconda/bin/perl /home/u027/project/software/bcbio/anaconda/bin/vep" # points to ../share/ensembl-vep-97.3-0/vep
REFERENCE_GENOME=/home/u027/project/data/reference/hg38.fa
echo "PROJECT_ID = ${PROJECT_ID}" # this the the folder (${BASE}/${PROJECT_ID}) where the downstream analysis will be done
echo "SOURCE_DIR = ${SOURCE_DIR}" # the command-line argument SOURCE_DIR is the general path to the source BAM files (VCF and PED already copied)
# enable running singletons
if [ -z $PBS_ARRAY_INDEX ]
then
if [ -z $INDEX ]
then
export PBS_ARRAY_INDEX=1
else
export PBS_ARRAY_INDEX=$INDEX
fi
fi
# change to the LOG folder
cd ${LOG_DIR}
################################
##### for each family ####
################################
FAMILY_ID=`head -n ${PBS_ARRAY_INDEX} ${FAMILY_IDS} | tail -n 1`
############################################################
### DNU and clean the each family VCF ###
### format: ${FAMILY_ID}-gatk-haplotype-annotated.vcf.gz ###
############################################################
echo "Performing DNU and cleaning of the ${FAMILY_ID}'s VCF file..."
time ${VT} decompose -s ${VCF_DIR}/${FAMILY_ID}-gatk-haplotype-annotated.vcf.gz -o ${VCF_DIR}/${FAMILY_ID}.decomp.vcf.gz
time ${VT} normalize ${VCF_DIR}/${FAMILY_ID}.decomp.vcf.gz -r ${REFERENCE_GENOME} -o ${VCF_DIR}/${FAMILY_ID}.norm.vcf.gz
time ${VT} uniq ${VCF_DIR}/${FAMILY_ID}.norm.vcf.gz -o ${VCF_DIR}/${FAMILY_ID}.DNU.vcf.gz
# remove sites with AC=0
time ${BCFTOOLS} view --min-ac=1 --no-update ${VCF_DIR}/${FAMILY_ID}.DNU.vcf.gz > ${VCF_DIR}/${FAMILY_ID}.AC0.vcf
# reset GT to no-call if num_ALT < num_ALT_THERSH or VAF < VAF_THRESH and GT != 0/0
# exlude variants from the blacklist (matching on chr,pos,ref,alt)
time ${PYTHON2} ${SCRIPTS_DIR}/filter_LQ_GT.py ${BLACKLIST} ${VCF_DIR}/${FAMILY_ID}.AC0.vcf ${VCF_DIR}/${FAMILY_ID}.clean.vcf
# bgzip and tabix it
time cat ${VCF_DIR}/${FAMILY_ID}.clean.vcf | ${BGZIP} > ${VCF_DIR}/${FAMILY_ID}.ready.vcf.gz
time ${TABIX} -p vcf ${VCF_DIR}/${FAMILY_ID}.ready.vcf.gz
# delete intermediate files
rm ${VCF_DIR}/${FAMILY_ID}-gatk-haplotype-annotated.vcf.gz
rm ${VCF_DIR}/${FAMILY_ID}.decomp.vcf.gz*
rm ${VCF_DIR}/${FAMILY_ID}.norm.vcf.gz*
rm ${VCF_DIR}/${FAMILY_ID}.DNU.vcf.gz*
rm ${VCF_DIR}/${FAMILY_ID}.AC0.vcf
# to avoid bgzip pipe broken annoying, but not problematic message - skip the next step
##############rm ${VCF_DIR}/${FAMILY_ID}.clean.vcf # to avoid bgzip pipe broken annoying (not problematic) message
echo ""
echo ""
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo "DNU, AC=0, num_ALT & VAF & blacklist cleaning of the ${FAMILY_ID}'s VCF file: done"
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo ""
echo ""
###########################################################
### run G2P for each family VCF (DD genes) ###
### format: ${FAMILY_ID}.ready.vcf.gz ###
###########################################################
echo "Performing G2P analysis (DD genes)for FAMILY_ID = ${FAMILY_ID}..."
######################################################################################IN_FILE=${VCF_DIR}/${FAMILY_ID}.ready.vcf.gz
IN_FILE=${VCF_DIR}/${FAMILY_ID}.clean.vcf #################################################################################
G2P_LOG_DIR=${G2P_DIR}/${FAMILY_ID}_LOG_DIR
mkdir ${G2P_LOG_DIR}
TXT_OUT=${G2P_LOG_DIR}/${FAMILY_ID}.report.txt
HTML_OUT=${G2P_LOG_DIR}/${FAMILY_ID}.report.html
VCF_KEYS='gnomADe|gnomADg'
time ${VEP} \
-i ${IN_FILE} \
--output_file ${G2P_LOG_DIR}/${FAMILY_ID}_inter_out.txt \
--force_overwrite \
--assembly GRCh38 \
--fasta ${REFERENCE_GENOME} \
--offline \
--merged \
--use_given_ref \
--cache --cache_version 97 \
--dir_cache /home/u027/project/software/bcbio/genomes/Hsapiens/hg38/vep \
--individual all \
--transcript_filter "gene_symbol in /home/u027/project/resources/genes_in_DDG2P.19092019.txt" \
--dir_plugins /home/u027/project/software/bcbio/anaconda/share/ensembl-vep-97.3-0 \
--plugin G2P,file='/home/u027/project/resources/DDG2P.19092019.csv',af_from_vcf=1,confidence_levels='confirmed&probable&both RD and IF',af_from_vcf_keys=${VCF_KEYS},log_dir=${G2P_LOG_DIR},txt_report=${TXT_OUT},html_report=${HTML_OUT}
echo ""
echo ""
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo "G2P analysis of FAMILY_ID = ${FAMILY_ID}: done"
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo ""
echo ""
###########################################################
### run VASE for each family VCF (de novo) ###
### format: ${FAMILY_ID}.ready.vcf.gz ###
###########################################################
echo "Performing de novo analysis with VASE for FAMILY_ID = ${FAMILY_ID} ..."
IN_FILE=${VCF_DIR}/${FAMILY_ID}.ready.vcf.gz
OUT_FILE=${VASE_DIR}/${FAMILY_ID}.strict.denovo.vcf
# PED_FILE=${PED_DIR}/${BATCH}_${FAMILY_ID}.ped
PED_FILE=${PED_DIR}/*_${FAMILY_ID}.ped
time ${VASE} \
-i ${IN_FILE} \
-o ${OUT_FILE} \
--log_progress \
--prog_interval 100000 \
--freq 0.0001 \
--gq 30 --dp 10 \
--het_ab 0.3 \
--max_alt_alleles 1 \
--csq all \
--biotypes all \
--control_gq 15 --control_dp 5 \
--control_het_ab 0.01 \
--control_max_ref_ab 0.05 \
--de_novo \
--ped ${PED_FILE}
# do some filtering on the denovo VCFs - exclude variants not on the 24 chr, as well as variants in LCR and telomere/centromere regions
### actually, ignore the filtering of variants in LCR and telomere/centromere regions --> more variants with “Unknown” status may be classified as “denovo” if enough support
cd ${VASE_DIR}
# index the denovo VCF
time ${GATK4} IndexFeatureFile -F ${OUT_FILE}
# select only variants on the 24 chromosomes
time ${GATK4} SelectVariants -R ${REFERENCE_GENOME} -V ${OUT_FILE} -O ${FAMILY_ID}.strict.24chr.denovo.vcf -L /home/u027/project/resources/24_chr.list --exclude-non-variants
# sort the VCF (maybe not needed?, but just in case, and it is quick)
rm ${FAMILY_ID}.strict.24chr.sort.denovo.vcf
grep '^#' ${FAMILY_ID}.strict.24chr.denovo.vcf > ${FAMILY_ID}.strict.24chr.sort.denovo.vcf \
&& grep -v '^#' ${FAMILY_ID}.strict.24chr.denovo.vcf | LC_ALL=C sort -t $'\t' -k1,1V -k2,2n >> ${FAMILY_ID}.strict.24chr.sort.denovo.vcf
# index the sorted VCF
time ${GATK4} IndexFeatureFile -F ${FAMILY_ID}.strict.24chr.sort.denovo.vcf
########################################################################################################################################
#### remove variants from LCR and telo-/centro-mere regions
###time ${GATK4} SelectVariants -R ${REFERENCE_GENOME} -V ${FAMILY_ID}.strict.24chr.sort.denovo.vcf -O ${FAMILY_ID}.clean.denovo.vcf \
###-XL /home/u027/project/resources/LCR.bed -XL /home/u027/project/resources/sv_repeat_telomere_centromere.bed --exclude-non-variants
#### split multi-allelic sites [by -m -any]
#### left-alignment and normalization [by adding the -f]
###file=${FAMILY_ID}.clean.denovo.vcf
###echo "$file"
###${BCFTOOLS} norm -f ${REFERENCE_GENOME} -m -any -Ov -o ${file/.clean.denovo.vcf/.ready.denovo.vcf} $file
#############################################################################################################################################
# split multi-allelic sites [by -m -any]
# left-alignment and normalization [by adding the -f]
file=${FAMILY_ID}.strict.24chr.sort.denovo.vcf
echo "$file"
${BCFTOOLS} norm -f ${REFERENCE_GENOME} -m -any -Ov -o ${file/.strict.24chr.sort.denovo.vcf/.ready.denovo.vcf} $file
# clean intermediate denovo files
rm ${FAMILY_ID}.strict.denovo.vcf*
rm ${FAMILY_ID}.strict.24chr.denovo.vcf*
rm ${FAMILY_ID}.strict.24chr.sort.denovo.vcf*
###rm ${FAMILY_ID}.clean.denovo.vcf*
# change back to the LOG folder
cd ${LOG_DIR}
echo ""
echo ""
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo "De novo analysis of FAMILY_ID = ${FAMILY_ID}: done"
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo ""
echo ""
#####################################################################
### run coverage for each proband (DD genes) ###
### format: ${SOURCE_DIR}/{PROBAND_ID}/{PROBAND_ID}-ready.bam ###
#####################################################################
#################################
##### for each proband ####
#################################
PROBAND_ID=`head -n ${PBS_ARRAY_INDEX} ${CHILD_IDS} | tail -n 1`
echo "Performing coverage analysis for PROBAND_ID = ${PROBAND_ID} ..."
BAM_FILE=${SOURCE_DIR}/${PROBAND_ID}/${PROBAND_ID}-ready.bam
OUT_FILE=${COV_DIR}/${PROBAND_ID}.DD15
time java -Xmx8g -jar ${GATK3} -T DepthOfCoverage -R ${REFERENCE_GENOME} -o ${OUT_FILE} -I ${BAM_FILE} -L ${TARGETS} \
--omitDepthOutputAtEachBase \
--minBaseQuality 20 \
--minMappingQuality 20 \
-ct 20 \
-jdk_deflater \
-jdk_inflater \
--allow_potentially_misencoded_quality_scores
echo ""
echo ""
echo "---------------------------------------------------------------------------------------"
echo "percentage of DD exons (+/-15) covered at least 20x in PROBAND_ID = ${PROBAND_ID} ..."
cat ${COV_DIR}/${PROBAND_ID}.DD15.sample_summary | awk '{print $7}'
echo "---------------------------------------------------------------------------------------"
# now compute the coverage per DD exon (+/-15bp) interval, adding the number of P/LP ClinVar variants (assertion criteria provided) in each interval
time ${PYTHON2} ${SCRIPTS_DIR}/generate_coverage_result_file.py ${COV_DIR}/${PROBAND_ID}.DD15.sample_interval_summary ${CLINVAR} ${COV_DIR}/${PROBAND_ID}.DD15.COV.txt
echo ""
echo ""
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo "Coverage analysis of PROBAND_ID = ${PROBAND_ID}: done"
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo ""
echo ""
#######################################################################
### for each proband generate the DECIPHER file ###
### ${VCF_DIR}/${FAMILY_ID}.ready.vcf.gz - the cleaned family VCF ###
### ${VASE_DIR}/${FAMILY_ID}.ready.denovo.vcf - the VASE file ###
### ${TRANS_MAP} - the current transcript mapping file ###
#######################################################################
echo "Generating the DECIPHER file for PROBAND_ID = ${PROBAND_ID} ..."
# first, split the family VCF to individual VCFs
# -c1: minimum allele count (INFO/AC) of sites to be printed
# split multi-allelic sites (by -m -any)
# left-alignment and normalization (by adding the -f)
file=${VCF_DIR}/${FAMILY_ID}.ready.vcf.gz
echo "splitting $file"
for indi in `${BCFTOOLS} query -l $file`; do
${BCFTOOLS} view -c1 -Oz -s $indi -o ${file/.vcf*/.$indi.rough.vcf.gz} $file
${BCFTOOLS} norm -f ${REFERENCE_GENOME} -m -any -Oz -o ${file/.vcf*/.$indi.vcf.gz} ${file/.vcf*/.$indi.rough.vcf.gz}
rm ${file/.vcf*/.$indi.rough.vcf.gz}
done
# VASE file - already split, left-aligned and normalized
# call the py scrpit
PED_FILE=${PED_DIR}/*_${FAMILY_ID}.ped
time ${PYTHON2} ${SCRIPTS_DIR}/generate_DEC_IGV.py ${PED_FILE} ${G2P_DIR}/${FAMILY_ID}_LOG_DIR/${FAMILY_ID}.report.txt \
${VASE_DIR}/${FAMILY_ID}.ready.denovo.vcf ${VCF_DIR}/${FAMILY_ID} ${DEC_DIR} ${IGV_DIR} ${IGV_DIR}/${FAMILY_ID} ${SOURCE_DIR} ${FAMILY_ID} ${WORK_DIR} ${TRANS_MAP}
echo ""
echo ""
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo "DECIPHER analysis of PROBAND_ID = ${PROBAND_ID}: done"
echo "^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
echo ""
echo ""
rm ${VCF_DIR}/${FAMILY_ID}.clean.vcf ##################################################