ILRI Research Computing

User Tools

Site Tools

You are here: start » mkatari-bioinformatics-august-2013-gatknotes
mkatari-bioinformatics-august-2013-gatknotes

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
Next revisionBoth sides next revision
mkatari-bioinformatics-august-2013-gatknotes [2014/06/11 12:52] – mkatarimkatari-bioinformatics-august-2013-gatknotes [2016/08/10 20:45] – mkatari
Line 29: Line 29:
   * create a bam index for dedup bam files   * create a bam index for dedup bam files
   * realign samples   * realign samples
 +
 +Once they are all processed
   * merge all samples to one bam file   * merge all samples to one bam file
   * sort and index this merged bam file   * sort and index this merged bam file
Line 34: Line 36:
  
 <code> <code>
-bowtie2 -x PTC_Human -U Cohen.fastq -S Cohen.sam +bowtie2 -x PTC_Human -U Sample1.fastq -S Sample1.sam 
-samtools view -bS Cohen.sam > Cohen.bam +samtools view -bS Sample1.sam > Sample1.bam 
-bowtie2 -x PTC_Human -U Sherman.fastq -S Sherman.sam + 
-samtools view -bS Sherman.sam > Sherman.bam+bowtie2 -x PTC_Human -U Sample2.fastq -S Sample2.sam 
 +samtools view -bS Sample2.sam > Sample2.bam 
 + 
 +bowtie2 -x PTC_Human -U Sample3.fastq -S Sample3.sam 
 +samtools view -bS Sample3.sam > Sample3.bam 
 + 
 +bowtie2 -x PTC_Human -U Sample4.fastq -S Sample4.sam 
 +samtools view -bS Sample4.sam > Sample4.bam
 </code> </code>
  
-The picard method to sort is preferred by GATK+The picard method to sort is preferred by GATK. In some cases PICARD uses the temp directory to do its sorting. You may run into an error that complains about running out of space. To avoid this problem simply create your own tmp directory and tell java that it should use it. See details [[https://www.biostars.org/p/42613/|here]]. 
 <code> <code>
-java -jar /export/apps/picard-tools/1.112/SortSam.jar \ +module load picard/1.133 
-   INPUT=Cohen.bam \ + 
-   OUTPUT=Cohen.sorted.bam \ +picard SortSam INPUT=Sample1.bam OUTPUT=Sample1.sorted.bam \ 
-   SORT_ORDER=coordinate +    SORT_ORDER=coordinate 
-    + 
-java -jar /export/apps/picard-tools/1.112/SortSam.jar \ +picard SortSam INPUT=Sample2.bam OUTPUT=Sample2.sorted.bam \ 
-   INPUT=Sherman.bam \ +    SORT_ORDER=coordinate 
-   OUTPUT=Sherman.sorted.bam \ + 
-   SORT_ORDER=coordinate+picard SortSam INPUT=Sample3.bam OUTPUT=Sample3.sorted.bam \ 
 +    SORT_ORDER=coordinate 
 + 
 +picard SortSam INPUT=Sample4.bam OUTPUT=Sample4.sorted.bam \ 
 +    SORT_ORDER=coordinate 
 + 
 </code> </code>
  
Line 56: Line 72:
  
 <code> <code>
-java -jar /export/apps/picard-tools/1.112/AddOrReplaceReadGroups.jar \ +picard AddOrReplaceReadGroups \ 
-   INPUT=Sherman.sorted.bam \ +   INPUT=Sample1.sorted.bam \ 
-   OUTPUT=ShermanRG.bam \ +   OUTPUT=Sample1RG.bam \ 
-   RGLB=Sherman \+   RGLB=Sample1 \
    RGPL=IonTorrent \    RGPL=IonTorrent \
    RGPU=None \    RGPU=None \
-   RGSM=Sherman+   RGSM=Sample1
  
-java -jar /export/apps/picard-tools/1.112/AddOrReplaceReadGroups.jar \ +picard AddOrReplaceReadGroups \ 
-   INPUT=Cohen.sorted.bam \ +   INPUT=Sample2.sorted.bam \ 
-   OUTPUT=CohenRG.bam \ +   OUTPUT=Sample2RG.bam \ 
-   RGLB=Cohen \+   RGLB=Sample2 \
    RGPL=IonTorrent \    RGPL=IonTorrent \
    RGPU=None \    RGPU=None \
-   RGSM=Cohen+   RGSM=Sample2 
 + 
 +picard AddOrReplaceReadGroups \ 
 +   INPUT=Sample3.sorted.bam \ 
 +   OUTPUT=Sample3RG.bam \ 
 +   RGLB=Sample3 \ 
 +   RGPL=IonTorrent \ 
 +   RGPU=None \ 
 +   RGSM=Sample3 
 + 
 +picard AddOrReplaceReadGroups \ 
 +   INPUT=Sample4.sorted.bam \ 
 +   OUTPUT=Sample4RG.bam \ 
 +   RGLB=Sample4 \ 
 +   RGPL=IonTorrent \ 
 +   RGPU=None \ 
 +   RGSM=Sample4 
 + 
 </code> </code>
  
 This will remove any reads that map to the same exact place. It is helpful to get rid of artifacts.  This will remove any reads that map to the same exact place. It is helpful to get rid of artifacts. 
 <code> <code>
-java -jar /export/apps/picard-tools/1.112/MarkDuplicates.jar \ + 
-   INPUT=CohenRG.bam \ +picard MarkDuplicates \ 
-   OUTPUT=Cohen.dedup.bam \ +   INPUT=Sample1RG.bam \ 
-   METRICS_FILE=Cohen.dedup.metrics \+   OUTPUT=Sample1.dedup.bam \ 
 +   METRICS_FILE=Sample1.dedup.metrics \
    REMOVE_DUPLICATES=TRUE \    REMOVE_DUPLICATES=TRUE \
-   ASSUME_SORTED=TRUE+   ASSUME_SORTED=TRUE \ 
 +   MAX_FILE_HANDLES_FOR_READ_ENDS_MAP=1000
  
-java -jar /export/apps/picard-tools/1.112/MarkDuplicates.jar \ + 
-   INPUT=ShermanRG.bam \ +picard MarkDuplicates \ 
-   OUTPUT=Sherman.dedup.bam \ +   INPUT=Sample2RG.bam \ 
-   METRICS_FILE=Sherman.dedup.metrics \+   OUTPUT=Sample2.dedup.bam \ 
 +   METRICS_FILE=Sample2.dedup.metrics \
    REMOVE_DUPLICATES=TRUE \    REMOVE_DUPLICATES=TRUE \
-   ASSUME_SORTED=TRUE+   ASSUME_SORTED=TRUE \ 
 +   MAX_FILE_HANDLES_FOR_READ_ENDS_MAP=1000 
 + 
 +picard MarkDuplicates \ 
 +   INPUT=Sample3RG.bam \ 
 +   OUTPUT=Sample3.dedup.bam \ 
 +   METRICS_FILE=Sample3.dedup.metrics \ 
 +   REMOVE_DUPLICATES=TRUE \ 
 +   ASSUME_SORTED=TRUE \ 
 +   MAX_FILE_HANDLES_FOR_READ_ENDS_MAP=1000 
 + 
 +picard MarkDuplicates \ 
 +   INPUT=Sample4RG.bam \ 
 +   OUTPUT=Sample4.dedup.bam \ 
 +   METRICS_FILE=Sample4.dedup.metrics \ 
 +   REMOVE_DUPLICATES=TRUE \ 
 +   ASSUME_SORTED=TRUE \ 
 +   MAX_FILE_HANDLES_FOR_READ_ENDS_MAP=1000 
 </code> </code>
  
 Index the files and realign them Index the files and realign them
 <code> <code>
-samtools index Cohen.dedup.bam  +samtools index Sample1.dedup.bam  
-samtools index Sherman.dedup.bam +samtools index Sample2.dedup.bam  
 +samtools index Sample3.dedup.bam  
 +samtools index Sample4.dedup.bam 
  
 #identifying indels #identifying indels
-java -Xmx2g -jar /export/apps/GenomeAnalysisTK/GenomeAnalysisTK-2.3-9-ge5ebf34/GenomeAnalysisTK.jar \+ 
 +module load gatk/3.3.0 
 + 
 +GenomeAnalysisTK \
    -T RealignerTargetCreator \    -T RealignerTargetCreator \
    -R PTC_Human.fasta \    -R PTC_Human.fasta \
-   -I Cohen.dedup.bam \ +   -I Sample1.dedup.bam \ 
-   -o CohenforIndelRealigner.intervals+   -o Sample1forIndelRealigner.intervals
    
- #identifying indels 
-java -Xmx2g -jar /export/apps/GenomeAnalysisTK/GenomeAnalysisTK-2.3-9-ge5ebf34/GenomeAnalysisTK.jar \ 
-   -T RealignerTargetCreator \ 
-   -R PTC_Human.fasta \ 
-   -I Sherman.dedup.bam \ 
-   -o ShermanforIndelRealigner.intervals 
  
-  +GenomeAnalysisTK \
- java -Xmx4g -jar /export/apps/GenomeAnalysisTK/GenomeAnalysisTK-2.3-9-ge5ebf34/GenomeAnalysisTK.jar \+
    -T IndelRealigner \    -T IndelRealigner \
    -R PTC_Human.fasta \    -R PTC_Human.fasta \
-   -I Cohen.dedup.bam \ +   -I Sample1.dedup.bam \ 
-  -targetIntervals CohenforIndelRealigner.intervals \ +   -targetIntervals Sample1forIndelRealigner.intervals \ 
-   -o Cohen.dedup.realign.bam+   -o Sample1.dedup.realign.bam
  
- java -Xmx4g -jar /export/apps/GenomeAnalysisTK/GenomeAnalysisTK-2.3-9-ge5ebf34/GenomeAnalysisTK.jar \ +</code>
-   -T IndelRealigner \ +
-   -R PTC_Human.fasta \ +
-   -I Sherman.dedup.bam \ +
-  -targetIntervals ShermanforIndelRealigner.intervals \ +
-   -o Sherman.dedup.realign.bam+
  
 +In some cases there may be a need to clean the sam/bam file(s) (soft-trimming the coordinates). To do this use CleanSam in Picard tools. You may want to just do it to all to avoid the error in a workflow, but it may not be necessary.
 +
 +<code>
 +picard CleanSam \
 +   INPUT=Sample4.dedup.realign.bam \
 +   OUTPUT=Sample4.clean.dedup.realign.bam
 </code> </code>
  
-Now we merge the bam files and then sort and index them+Now we merge the bam files and then sort and index them. If you cleaned the bam file, remember to use the cleaned ones.
  
 <code> <code>
-java -jar /export/apps/picard-tools/1.112/MergeSamFiles.jar \ +picard MergeSamFiles \ 
-   INPUT=Sherman.dedup.realign.bam \ +   INPUT=Sample1.dedup.realign.bam \ 
-   INPUT=Cohen.dedup.realign.bam \ +   INPUT=Sample2.dedup.realign.bam \ 
-   OUTPUT=ShermanCohenMerged.bam  +   INPUT=Sample3.dedup.realign.bam \ 
 +   INPUT=Sample4.dedup.realign.bam \ 
 +   OUTPUT=AllMerged.bam  
  
-samtools sort ShermanCohenMerged.bam ShermanCohenMerged.sorted+picard SortSam INPUT=AllMerged.bam OUTPUT=AllMerged.sorted.bam SORT_ORDER=coordinate 
 + 
 +samtools index AllMerged.sorted.bam
  
-samtools index ShermanCohenMerged.sorted.bam  
 </code> </code>
  
  
-Finall !! run gatk+Finally !! run gatk
  
 <code> <code>
-java -jar /export/apps/GenomeAnalysisTK/GenomeAnalysisTK-2.3-9-ge5ebf34/GenomeAnalysisTK.jar \ +GenomeAnalysisTK -T UnifiedGenotyper \ 
-   -T UnifiedGenotyper \ +   -I AllMerged.sorted.bam \
-   -I ShermanCohenMerged.sorted.bam \+
    -R PTC_Human.fasta \    -R PTC_Human.fasta \
    --output_mode EMIT_VARIANTS_ONLY \    --output_mode EMIT_VARIANTS_ONLY \
Line 151: Line 207:
    -glm SNP \    -glm SNP \
    -o PTC_human.gatk.vcf    -o PTC_human.gatk.vcf
 +
 +</code>
 +
 +If you want to load the vcf file into IGV, remember to index it first.
 +<code>
 +module load igvtools
 +igvtools index PTC_human.gatk.vcf
 +</code>
 +
 +If you would like to generate a table of from the vcf file use the following command
 +<code>
 +GenomeAnalysisTK \
 +     -R PTC_Human.fasta
 +     -T VariantsToTable \
 +     -V PTC_human.gatk.vcf \
 +     -F CHROM -F POS -F ID -F QUAL -F AC \
 +     -GF GT -GF GQ \
 +     -o PTC_human.gatk.vcf.table
 +</code>
 +
 +In order to filter your vcf file based on quality measures, depth, and also statistical significance, you can use variant filter option in the gatk toolkit. Below is an example of suggested filters for data that has low coverage.
 +
 +<code>
 +GenomeAnalysisTK \
 +    -R PTC_Human.fasta \
 +    -T VariantFiltration \
 +    -o PTC_human.gatk.filter.vcf \
 +    --variant PTC_human.gatk.vcf \
 +    --filterExpression "QD < 2.0 || MQ < 40.0 || FS > 60.0 || HaplotypeScore >13.0" \
 +    --filterName "mannyfilter"
 +
 +</code>
 +
 +Good descriptions of the different information on vcf files [[https://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_gatk_tools_walkers_annotator_HaplotypeScore.php|GATK Docs]]
 +
 +Finally to save the SNPs that passed your filter, you simply use the selectvariant tool.
 +
 +<code>
 +
 +GenomeAnalysisTK \
 +    -T SelectVariants \
 +    --variant PTC_human.gatk.filter.vcf \
 +    -o PTC_human.gatk.filter.only.vcf \
 +    -ef \
 +    -R PTC_Human.fasta
 +
 </code> </code>
mkatari-bioinformatics-august-2013-gatknotes.txt · Last modified: 2016/08/17 08:37 by mkatari