some Mito-mutations can't detected out right from amplicon data using Mutect2
GATK commands: java -Xmx16g -Djava.io.tmpdir=./tmp -jar gatk-package-22.214.171.124-local.jar Mutect2 -R hs37d5.fa -L MT.bed --max-reads-per-alignment-start 0 --mitochondria-mode --kmer-size 20 -I MT.primerclipped.bam -O MT.output.vcf
Hi, I'm runing GATK4 mutect2 mitochondria-mode on MT amplicon data, but an obvious error occurred. The base site has high depth and good quality, but the result is None. Importantly, several mutation sites are both FN. I don't know why these mutation sites can't be detected. Is it because I lost a certain parameter? Or can Mutect2 not be used for amplicon data? Or is there a bug in the program?
For example, Mutect2 reports:
#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT D10_ID-20200112-ZD_V300045813_L02_1-65
MT 513 . GCA G . . DP=10178;ECNT=1;MBQ=37,37;MFRL=0,0;MMQ=60,60;MPOS=122;OCM=0;POPAF=2.40;RPA=5,4;RU=CA;STR;TLOD=15993.15 GT:AD:AF:DP:F1R2:F2R1:SB 0/1:314,9801:0.995:10115:198,5990:59,3260:100,214,3767,6034
MT 16217 . T C . . DP=1615;ECNT=1;MBQ=37,37;MFRL=0,0;MMQ=60,60;MPOS=20;OCM=0;POPAF=2.40;TLOD=67.15 GT:AD:AF:DP:F1R2:F2R1:SB 0/1:1580,35:0.022:1615:0,0:1569,35:1580,0,35,0
IGV gives me for the BAM:
A few questions for you:
- How are you processing your data? Have you looked at the pipeline available inn Terra: https://app.terra.bio/#workspaces/help-gatk/Mitochondria-SNPs-Indels-hg38
- If you running it the same way as shown in Terra, then please share your data with us for troubleshooting. Please share your bam, bed and ref-fasta files using these instructions: https://gatk.zendesk.com/hc/en-us/articles/360035889671
@rime We are working on some issues involving dangling ends in the assembly graph that are especially pervasive with amplicon data. The new -linked-de-bruijn-graph argument has been seen to help in some cases and I would recommend it with all amplicon data.
Please sign in to leave a comment.