終於看到了一個完整的mutect2使用指令碼

因為嫌麻煩，所以一直使用的是簡化版mutect2流程，其實就一個命令：

time $GATK  --java-options "-Xmx10G -Djava.io.tmpdir=./"  Mutect2 -R $reference \
-I $tumor_bam  -tumor $(basename "$tumor_bam" _recal.bam) \
-I $normal_bam -normal $(basename "$normal_bam" _recal.bam) \
-O ${sample}_mutect2.vcf
$GATK  FilterMutectCalls -V ${sample}_mutect2.vcf -O ${sample}_somatic.vcf

但是很明顯，這其實不符合官網教程的理念

https://gatkforums.broadinstitute.org/gatk/discussion/9183/how-to-call-somatic-snvs-and-indels-using-mutect2
https://software.broadinstitute.org/gatk/documentation/article?id=9183

但是官網教程的確太繁瑣，裡面涉及到6個步驟，而我只是執行了最後一個！

因為種種原因，沒能抽出時間細緻的探索mutect2的用法，但是無意中搜尋到指令碼一個： https://figshare.com/articles/scripts_sh/4542397

可以說是非常良心了：

#!/bin/sh

##GATK bundle download
wget ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/hg19/ucsc.hg19.fasta.gz -O /path/to/GATK/bundle/ucsc.hg19.fasta.gz
wget ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/hg19/dbsnp_138.hg19.vcf.gz -O /path/to/GATK/bundle/dbsnp_138.hg19.vcf.gz
wget https://ndownloader.figshare.com/files/7354246 -O /path/to/GATK/bundle/TP53.sorted.bed
wget https://ndownloader.figshare.com/files/7354213 -O /path/to/GATK/bundle/CosmicCodingMuts.chr.sort.head.vcf

##ANNOVAR database files download
export PATH=$PATH:/path/to/annovar

annotate_variation.pl -buildver hg19 -downdb -webfrom annovar refGene humandb/
annotate_variation.pl -buildver hg19 -downdb cytoBand humandb/
annotate_variation.pl -buildver hg19 -downdb genomicSuperDups humandb/ 
annotate_variation.pl -buildver hg19 -downdb -webfrom annovar esp6500siv2_all humandb/
annotate_variation.pl -buildver hg19 -downdb -webfrom annovar 1000g2015aug humandb/
annotate_variation.pl -buildver hg19 -downdb -webfrom annovar snp138 humandb/
annotate_variation.pl -buildver hg19 -downdb -webfrom annovar dbnsfp30a humandb/
annotate_variation.pl -buildver hg19 -downdb -webfrom annovar cosmic70 humandb/
annotate_variation.pl -buildver hg19 -downdb -webfrom annovar exac03 humandb/ 
annotate_variation.pl -buildver hg19 -downdb -webfrom annovar clinvar_20160302 humandb/



#define the reference file path
GATK=/path/to/GATK/GenomeAnalysisTK.jar
REF=/path/to/GATK/bundle/ucsc.hg19.fasta
DBSNP=/path/to/GATK/bundle/dbsnp_138.hg19.vcf
COSMIC=/path/to/GATK/bundle/CosmicCodingMuts.chr.sort.head.vcf
BED=/path/to/GATK/bundle/TP53.sorted.bed

#create a Panel of Normals (PoN) vcf file from the 4 low-grade tumour samples
for pathandfile in /path/to/EGA/normal/*.clean.recal.TP53.bam ; do
    basewithpath=${pathandfile%.clean.recal.TP53.*}
    basenopath=$(basename $basewithpath)

java -jar $GATK \
    -T MuTect2 \
    -R $REF \
    -I:tumor $(echo $basewithpath).clean.recal.TP53.bam \
    --dbsnp $DBSNP \
    --cosmic $COSMIC \
    --artifact_detection_mode \
    -L $BED \
    -o $(echo $basewithpath).clean.recal.TP53.normal.vcf
done


java -jar $GATK \
    -T CombineVariants \
    -R $REF \
    -V /path/to/EGA/normal/GB544-10_S18.clean.recal.TP53.normal.vcf -V /path/to/EGA/normal/GB624-11_S81.clean.recal.TP53.normal.vcf -V /path/to/EGA/normal/GB730-12_S41.clean.recal.TP53.normal.vcf -V /path/to/EGA/normal/GB909-13_S90.clean.recal.TP53.normal.vcf \
    -minN 2 \
    --setKey "null" \
    --filteredAreUncalled \
    --filteredrecordsmergetype KEEP_IF_ANY_UNFILTERED \
    -L $BED \
    -o /path/to/EGA/normal/MuTect2_PON.vcf
    
    
    

#call somatic variants
for pathandfile in /path/to/EGA/tumor/*.clean.recal.TP53.bam ; do
    basewithpath=${pathandfile%.clean.recal.TP53.*}
    basenopath=$(basename $basewithpath)
    
java -jar $GATK \
    -T MuTect2 \
    -R $REF \
    --dbsnp $DBSNP \
    --cosmic $COSMIC \
    -dt NONE \
    --input_file:tumor $(echo $basewithpath).clean.recal.TP53.bam \
    --intervals $BED \
    -PON /path/to/EGA/normal/MuTect2_PON.vcf \
    -o $(echo $basewithpath).clean.recal.TP53.vcf

done




#ANNOVAR annotation
mkdir /path/to/EGA/tumor/ANNOVAR
cp /path/to/EGA/tumor/*.vcf /path/to/EGA/tumor/ANNOVAR

##convert file format
for pathandfile in /path/to/EGA/tumor/ANNOVAR/*.vcf ; do
    filename=${pathandfile%.*} 
    convert2annovar.pl --format vcf4 --includeinfo --withzyg $pathandfile > $(echo $filename).avinput
done

##add sample information
for pathandfile in /path/to/EGA/tumor/ANNOVAR/*.avinput ; do 
    filename=$(basename $pathandfile)
    mainfilename=${filename%.clean.recal.TP53.avinput}
    pathandmain=${pathandfile%.*}
    awk -v var1="$mainfilename" '{OFS = "\t" ; print $0, var1}' $pathandfile > $(echo $pathandmain).avinputs
done

##merge file
cat /path/to/EGA/tumor/ANNOVAR/*.avinputs > /path/to/EGA/tumor/ANNOVAR/TP53.Tonly.avinput
    
##annotate
table_annovar.pl /path/to/EGA/tumor/ANNOVAR/TP53.Tonly.avinput /path/to/annovar/humandb/ -buildver hg19 -out /path/to/EGA/tumor/ANNOVAR/TP53.Tonly -remove -protocol refGene,cytoBand,genomicSuperDups,esp6500siv2_all,1000g2015aug_all,snp138,dbnsfp30a,cosmic70,exac03,clinvar_20160302 -operation g,r,r,f,f,f,f,f,f,f --otherinfo

雖然這個流程是基於 hg19 參考基因組的，但是很容易就能改寫為 hg38版本的！

還有一個問題是他那個流程的專案背景是，得到的腫瘤樣品是沒有配對normal的，而是 create a Panel of Normals (PoN) vcf file from the 4 low-grade tumour samples

而且，這個流程是基於 GATK3成熟版本的，並不是GATK4哦。

如果大家有GATK4的MUTECT2經驗，可以交流一下哈。