How do I ensure sequence context-dependent artifacts, such as for FFPE deamination, are filtered using the current version of GATK?
Dear GATK Team,
In previous (How to) Call somatic mutations using GATK4 Mutect2 guidance, it was possible to run the tool FilterByOrientationBias (EXPERIMENTAL) and an 'orientation bias' filter was added to account for 'sequence context-dependent artifacts, e.g. OxoG or FFPE deamination'.
However, this tool is now not available in more recent versions of GATK. Although an orientation bias filter is still applied with FilterMutectCalls, it seems to have a different definition in the VCF compared to that applied with FilterByOrientationBias (EXPERIMENTAL).
Definitions for orientation bias in the FILTER ID field:
- FilterByOrientationBias (EXPERIMENTAL) = "Orientation bias (in one of the specified artifact mode(s) or complement) seen in one or more samples."
- FilterMutectCalls = "orientation bias detected by the orientation bias mixture model."
How do I ensure sequence context-dependent artifacts, such as for FFPE deamination, are filtered using the current version of GATK?
The current (How to) Call somatic mutations using GATK4 Mutect2 guidance, in the description for the Mutect2 Read Orientation Artifacts Workflow, says that Mutect2's orientation bias filter should be used for FFPE tumour samples but a specific reason related to sequence context-dependent artifacts is not given. In addition, there is no reference to CollectSequencingArtifactMetrics (Picard) or FilterByOrientationBias (EXPERIMENTAL) as previous tools and therefore, I do not know if the Mutect2 Read Orientation Artifacts Workflow specifically filters for OxoG and FFPE deamination artifacts.
Thank you for your time and help.
Kind regards.
-
I'm having similar problems. I have a high-proportion of G>T and C>A calls in my vcf even after filtering and specifying ob-priors
-
Hi ISmolicz,
The -f1r2-tar-gz output of Mutect2 in more recent GATK versions replaces CollectSequencingArtifactMetrics. This file contains all of the information for orientation bias artifacts, including OxoG and FFPE deamination so there is no need to specify a specific artifact. The documentation is a little bit different because FilterMutectCalls uses a different model than FilterByOrientationBias but the filtering and output should be the same. Please let me know if you have any other questions or concerns.
Kind regards,
Pamela
Please sign in to leave a comment.
2 comments