Mapping and variant calling for hypermutated cancers
Dear all,
I am interested in the analysis of WES data for hypermutated cancers, like colorectal, lung, melanoma cancers... I mean cancers with >200 somatic mutations per Mb.
Until now, I have analysed paired data with low (<10/Mb) somatic mutations. Thus, I wonder if mapping and variant calling steps require more attention in this context.
The high mutation rate can make the mapping more complicated. For mapping WES data, I have used BWA-backtrack and BWA-MEM with default settings in the past. With very high mutation rate, should we keep only reads with very stringent mapping score (for unique alignment only)? Should we modify the default parameters?
For variant (SNV+short INDEL) calling, I am used to VarScan2. I would like to test an additional tool. MuTect (which detects SNV only) is one of the most frequently used variant callers for WES. Since it does assumption on mutation frequency:
From MuTect publication:
this is probably not an appropriate tool to use in this context, unless it is possible to change the range of expected somatic mutations.
What do you think about Strelka on WES data? It has the advantage to detect short INDEL too.
Looking forward to hear from your experience,
Thank you,
Jane
I am interested in the analysis of WES data for hypermutated cancers, like colorectal, lung, melanoma cancers... I mean cancers with >200 somatic mutations per Mb.
Until now, I have analysed paired data with low (<10/Mb) somatic mutations. Thus, I wonder if mapping and variant calling steps require more attention in this context.
The high mutation rate can make the mapping more complicated. For mapping WES data, I have used BWA-backtrack and BWA-MEM with default settings in the past. With very high mutation rate, should we keep only reads with very stringent mapping score (for unique alignment only)? Should we modify the default parameters?
For variant (SNV+short INDEL) calling, I am used to VarScan2. I would like to test an additional tool. MuTect (which detects SNV only) is one of the most frequently used variant callers for WES. Since it does assumption on mutation frequency:
From MuTect publication:
We used a fixed threshold of 6.3 for all results presented here unless indicated otherwise. This threshold corresponds to a 10^6.3:1 odds ratio in favor of the reference model, which is reasonable because the frequency of mutations in many tumors is only 1–10 per Mb and thus the a priori odds of a site harboring a mutation may be as low as 1:10^5 or 1:10^6.
What do you think about Strelka on WES data? It has the advantage to detect short INDEL too.
Looking forward to hear from your experience,
Thank you,
Jane
Comment