Mutect2 AF does not match AD and DP

Answered

Follow

yqiu

• March 07, 2022 23:03

We have one indel detected by Mutect2 but the AF reported does not match AD and DP. As shown below, AF is 0.425 and 0.134 for two alternative alleles. But AD is only 258 and 255 respectively with 18941 total reads.

CHROM    POS    .    GCCAA    GA,G    .    multiallelic    AS_FilterStatus=SITE|SITE;AS_SB_TABLE=9218,9210|149,109|180,75;DP=43718;ECNT=40;GERMQ=93;MBQ=20,20,20;MFRL=122,122,122;MMQ=60,60,60;MPOS=55,55;POPAF=7.3,7.3;TLOD=75.03,75.87    GT:AD:AF:DP:F1R2:F2R1:SB    0/1/2:18428,258,255:0.425,0.134:18941:18375,82,80:0,0,0:9218,9210,329,184

We expect AF to be AD divided by DP and not sure if there is some extra complication that we are no aware of. 

0

• 

  Genevieve Brandt (she/her)

  • March 08, 2022 18:36

  Hi yqiu,

  I am going to move your post into our Community Discussions -> General Discussion topic, as the Somatic topic is for reporting bugs and issues with GATK.

  You can read more about our forum guidelines and the topics here: Forum Guidelines.

  Best,

  Genevieve

  0

  Comment actions Permalink
• 

  Genevieve Brandt (she/her)

  • March 15, 2022 17:06

  Hi yqiu,

  There could be some uninformative reads here. Please see this article regarding informative reads: https://gatk.broadinstitute.org/hc/en-us/articles/360035532252-Allele-Depth-AD-is-lower-than-expected

  Let me know if you have any further questions.

  Best,

  Genevieve

  0

  Comment actions Permalink
• 

  yqiu

  • March 18, 2022 16:42

  Hi Genevieve,

  Thanks for the response and the link you shared is very useful.

  But it may not directly relate to my question. I get that there are some reads are uninformative and they are not counted in AD/DP. My question is why AF is not matching AD and DP. In the vcf record I shared, , AF is 0.425 and 0.134 for two alternative alleles. But AD is only 258 and 255 respectively with 18941 total reads. If AF=AD/DP, there should be 8,049 and 2,538 instead.

  Hope I explained it clear enough.

  0

  Comment actions Permalink
• 

  Pamela Bretscher

  • March 21, 2022 13:29

  Hi yqiu,

  I'm working on this post along with Genevieve while she is out of the office this week. From my understanding, I believe the discrepancy is coming from the fact that the DP value includes uninformative reads while the AF calculation only takes into account the DP of informative reads. Therefore, it looks like you have 18941 total reads but the calculation is using a much lower DP value for only the informative reads. Please let me know if this helps answer your question.

  Kind regards,

  Pamela

  0

  Comment actions Permalink

Please sign in to leave a comment.