Test-drive the GATK tools and Best Practices pipelines on Terra


Check out this blog post to learn how you can get started with GATK and try out the pipelines in preconfigured workspaces (with a user-friendly interface!) without having to install anything.

which is the best way to see ALL the nucleotides in EVERY position of a target?

Hi
I would like to ask you which is the best way to see ALL the nucleotides in EVERY position of a target. I used EMIT_ALL_SITES option with UnifiedGenotyper but the output miss an information: I obtain the genotype and the coverage but no information about other possible nucleotides present but not called (for example 148 G and 2 T with genotype 0/0 and DP 150). Can you help me?
Thank you in advance

Answers

  • SheilaSheila Broad InstituteMember, Broadie, Moderator admin

    @Marta‌

    Hi,

    You can try using GVCF mode in HaplotypeCaller. GVCF mode is used to record the presence of alternate alleles to a high level of sensitivity. However, if the fraction of a particular alternate allele present is too low, it will not be recorded. In your case, this will give you the information you want for sites that have even a small chance of being variant, but not for sites that are clearly errors (such as your example of 148G/2T above).

    If you absolutely need to know every nucleotide present per site in your reads, you can use Samtools pileup. GATK tools do not emit that information because it is not an efficient way of calling variants.

    -Sheila

  • MartaMarta Member

    Thank you. I am trying to check for mosaicism at very low percentage. Any other suggestion for that? I am working on a small gene panel (not exome) and the platform is Ion Torrent.Thanks

  • SheilaSheila Broad InstituteMember, Broadie, Moderator admin

    @Marta‌

    Hi,

    GATK callers are not designed for this type of experimental design. They are designed to detect germline mutations.

    You can look into MuTect which is designed to detect somatic mutations. MuTect is much more sensitive than GATK callers. Please read about MuTect here: http://www.broadinstitute.org/cancer/cga/mutect

    -Sheila

  • samtools mpileup perhaps? You could then parse the pileups yourself to get low-fraction variants. I wouldn't recommend MuTect, it was designed to call somatic variants, most of which are at higher frequencies than what I assume you're looking for.

Sign In or Register to comment.