UGP Variant Pipeline 0.0.3

Utah Genome Project

Sept. 2013

Data Source

Data sets used for the variant calling pipeline come from the Broad GSA (GATK) group as the 'GATK resource bundle 2.5' version 2.8

wget -r ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/2.5/b37

• Reference Genome (GRCh37):
  • human_g1k_v37_decoy.fasta

• VCF files for RealignerTargetCreator knowns and dbsnp for BaseRecalibrator.
  • known_indel: /data/GATK_Bundle/Mills_and_1000G_gold_standard.indels.b37.vcf
  • known_indel: /data/GATK_Bundle/1000G_phase1.indels.b37.vcf
  • known_dbsnp: /data/GATK_Bundle/dbsnp_137.b37.vcf

• Background BAMS used running UnifiedGenotyper
  • 1000Genomes sets:
  • CEU
  • GBR

• Resource files for VariantRecalibrator_SNP
  • hapmap_3.3.b37.vcf
  • 1000G_omni2.5.b37.vcf
  • 1000G_phase1.snps.high_confidence.b37.vcf

• Resource files for VariantRecalibrator_INDEL
  • Mills_and_1000G_gold_standard.indels.b37.vcf
  • 1000G_phase1.indels.b37.vcf

Sequencing

This pipeline is designed for 100 bp Illumina HiSeq PE exome or WGS sequence data with Sanger/Illumina 1.9 quality encoding, and uses Illumina naming convention found here [1]

Validate File Integrity with md5sum

An md5sum signature should be provided for each FastQ file by the sequencing center. After the file has been downloaded, locally check the md5sum to be sure that no data corruption occurred during the file transfer.

md5sum file.fastq > file_local.md5
diff file_local.md5 file_provided.md5

If the md5sum signature differs from that provided for the file:

• Check to be sure you have the correct file.
• Check if the md5sum was calculated on that compressed or uncompressed file by the provider and be sure to do the same with the local copy.
• Try the download again.
• Contact the sequence provider.

FastQ File Analyses

fastqc Sample1_L1_R1.txt

From the fastqc_data.txt file we check the following values:

• Encoding (must be Sanger / Illumina 1.9)
• Total Sequences (Need to develop a acceptable range)
• Filtered Sequences (Should be 0 or at least very low)
• Sequence length (must be ~ 100 bp)
• %GC (should be 45 < x < 55)
• Total Duplicate Percentage (This value may not be valuable - an acceptable range has not been determined).

Alignment

Align reads to the genome with bwa.

Index the reference sequence if this has not already been done

bwa index -a bwtsw human_g1k_v37_decoy.fasta

The 'bwa aln' program will find the reference coordinates of the input reads (independent of their mate-pair). The following parameters are those used by the 1KG project for aligning Illumina data.

bwa aln -q 15 human_g1k_v37_decoy.fasta file.fastq > Sample1_L1_R1.sai # One lane of reads first in pair
bwa aln -q 15 human_g1k_v37_decoy.fasta file.fastq > Sample1_L1_R2.sai # One lane of reads second in pair

The 'bwa sampe'. For paired-end reads, the maximum insert size is taken to be 3 times the expected insert size and read group added.

bwa sampe -r "read group" -P Sample1_L1_R1.sai Sample1_L1_R1.sai Sample1_L1_R1.fastq Sample1_L1_R2.fastq | samtools view -bSho BAM_FILE -

This will switch to bwa mem soon.

bwa mem -M -R "read group" human_g1k_v37_decoy.fasta Sample1_L1_R1.fq Sample1_L1_R2.fq | samtools view -bSho BAM_FILE - 

BAM File Analyses

Alignment BAM files are improved in various ways to help increase the quality and speed of subsequent variant calling steps.

Merge lane level BAMs to individual

java -Xmx10g -XX:ParallelGCThreads=10 -Djava.io.tmpdir=/tmp MergeSamFiles.jar
    INPUT=Sample1_L1.rg.bam                     
    INPUT=Sample1_L2.rg.bam                    
    OUTPUT=Sample1.bam                          
    VALIDATION_STRINGENCY: LENIENT
    MAX_RECORDS_IN_RAM: 5000000
    CREATE_INDEX: True
    SORT_ORDER: coordinate
    ASSUME_SORTED: True 
    MERGE_SEQUENCE_DICTIONARIES:
    USE_THREADING: True
    2> MergeSamFiles.report

Mark Duplicates

Remove PCR/Optical duplicate reads

java -Xmx10g -XX:ParallelGCThreads=10 -Djava.io.tmpdir=/tmp MarkDuplicates.jar
   INPUT=Sample1.bam                   
   OUTPUT=Sample1.dedup.bam            
   METRICS_FILE=lane1_dup_metrics.txt  
   VALIDATION_STRINGENCY: LENIENT
   COMPRESSION_LEVEL:
   MAX_RECORDS_IN_RAM: 5000000
   CREATE_INDEX: True
   ASSUME_SORTED: True       
   2> MarkDuplicates.log

Develop range for duplicate levels here

Local Realignment of Indels

java -Xmx10g -Djava.io.tmpdir=/tmp GenomeAnalysisTK.jar
   -I dedup_bam.list                                    
   -o Realign_Intervals.txt                            
   -T RealignerTargetCreator                           
   -R human_g1k_v37_decoy.fasta                              
   -known [list from above]
   -nt 24                                              
   2> RealignerTargetCreator.log

Base Quality Score Recalibration

java -Xmx10g -Djava.io.tmpdir=/tmp GenomeAnalysisTK.jar
  -T PrintReads                       
  -I Sample1.realign.bam              
  -o Sample1.recal.bam                
  -R human_g1k_v37_decoy.fasta              
  -BQSR recalibration_report.grp      

BAM Quality Control

CollectAlignmentSummaryMetrics

java -Xmx10g -XX:ParallelGCThreads=10 -Djava.io.tmpdir=/tmp CollectAlignmentSummaryMetrics.jar
   I=/output/filename.b37_1kg.sorted.bam                                      
   O=/output/filename.b37_1kg.AlignmentSummaryMetrics                         
   R= human_g1k_v37_decoy.fasta                                                    
   VALIDATION_STRINGENCY=LENIENT                                             
   PROGRAM: QualityScoreDistribution

ReduceReads

java -Xmx10g -XX:ParallelGCThreads=10 -Djava.io.tmpdir=/tmp GenomeAnalysisTK.jar
  -T ReduceReads
  -I Sample1.recal.bam
  -o Sample1.reduced.bam
  -R human_g1k_v37_decoy.fasta
  2> ReduceReads.report

Variant Calling

UnifiedGenotyper

SNP calling

java -Xmx10g -XX:ParallelGCThreads=10 -Djava.io.tmpdir=/tmp GenomeAnalysisTK.jar
  -T UnifiedGenotyper                  
  -I reduced_bam.list
  -I 1000G background_bam.list
  -o Sample1.raw.vcf                  
  -R human_g1K_v37.fasta
  -nt 4
  -nct 6
  -stand_call_conf 30.0
  -stand_emit_conf 30.0
  -glm SNP   
  -out_mode EMIT_VARIANTS_ONLY
  -L NCBI Ref_GRCh37 region.bed 
  2> UnifiedGenotyper.report

VariantRecalibrator

java -Xmx4g -jar GenomeAnalysisTK.jar \
  -T VariantRecalibrator              \
  -R human_g1K_v37.fasta              \
  -input Sample1.raw.vcf              \
  -resource:hapmap,known=false,training=true,truth=true,prior=15.0 hapmap_3.3.b37.vcf          \
  -resource:omni,known=false,training=true,truth=true,prior=12.0 1000G_omni2.5.b37.vcf        \
  -resource:dbsnp,known=true,training=false,truth=false,prior=6.0 dbsnp_137.b37.vcf            \

ADD 1KG

  -an QD -an HaplotypeScore -an MQRankSum -an ReadPosRankSum -an FS -an MQ -an InbreedingCoeff \
  -mode SNP                           \
  -recalFile Sample1.recal            \
  -tranchesFile Sample1.tranches      \
  -rscriptFile Sample1.plots.R        \
  2> VariantRecalibrator.error

ApplyRecalibration

java -Xmx4g -jar GenomeAnalysisTK.jar \
  -T ApplyRecalibration               \
  -input Sample1.raw.vcf              \
  -o Sample1.vqsr.vcf                 \
  -R human_g1K_v37.fasta              \
  --ts_filter_level 99.0                \
  -tranchesFile Sample1.tranches      \
  -recalFile Sample1.recal            \
  -mode SNP                           \
  2> ApplyRecalibration.error

Variant File QC

Quality Metrics on variants

• Ti/Tv Ratio (2.1 for WGS ~2.8 for exome)
• HapMap concordance
• SNV/Indel Counts
• Rare variant enrichment
• DP
• Q
• GQ