Notice:
If you happen to see a question you know the answer to, please do chime in and help your fellow community members. We appreciate your help!

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.

GATK4 - Available generic command line options and read filters for every tool ?

tonytony Member ✭✭
edited November 2018 in Ask the GATK team

Hi,

I am currently trying to set up a pipeline using Mutect2 (GATK 4.0.11.0). I would like to know exactly what this tool is doing with my reads and also whether I am able to supply a multi-threading option. I can not easily find information on that (command line help, online tool documentation or snippets in tutorials).

Is there a way to know for any tool :

- which read filters are applied by default 
- if there is a support for multi-threading 

?

Thank you very much,

:smile:

Anthony

Answers

  • bhanuGandhambhanuGandham Cambridge MAMember, Administrator, Broadie, Moderator admin

    Hi @tony

    Here is some information on multithreading.

    Details on default values and filters used in Mutect2 can be found in its doc page here. The doc pages can give you information on other tools as well.

    Regards
    Bhanu

  • tonytony Member ✭✭

    Thanks @bhanuGandham

    OK so -nct or -nt old options do not exist anymore in GATK4 and the only way to do some multi-threading is to use the Spark functionality when available. In particular Mutect2 does not support multi-threading as I understand it and one should better use a Scatter/Gather approach in this case, right ?

    Nevertheless, it is still unclear to me what are the default read filters applied by Mutect2. The options -RF and -DF are both left blank [] in the tool doc. Does Mutect2 apply no filter at all ? It would not seem in agreement with the fact that you advice to disable the MOSCO filter in some Mutect2 use cases. I suppose that this filter is then applied by default in Mutect2 but I can not find that information anywhere in the tool doc. Am I missing a section or something in the docs ?

    Many thanks and best wishes,
    Anthony

  • bhanuGandhambhanuGandham Cambridge MAMember, Administrator, Broadie, Moderator admin

    Hi @tony

    OK so -nct or -nt old options do not exist anymore in GATK4 and the only way to do some multi-threading is to use the Spark functionality when available. In particular Mutect2 does not support multi-threading as I understand it and one should better use a Scatter/Gather approach in this case, right ?

    Yes, this is correct!

    The options -RF and -DF are both left blank [] in the tool doc. Does Mutect2 apply no filter at all ? It would not seem in agreement with the fact that you advice to disable the MOSCO filter in some Mutect2 use cases.

    -RF and -Df are both left blank to allow the user to apply to express what reads they want to allow through. Mutect2 does apply default filters such as --disable-tool-default-read-filters is set to false, --use-filtered-reads-for-annotations is set to false, --max-suspicious-reads-per-alignment-start is set to 0. MOSCO option will keep only reads whose mate maps to the same contig or is unmapped. Disabling the MOSCO filter is suggested so that reads on alternate contigs and read pairs that span contigs can lend support to variant calls. For more information onread filters please follow these links:
    (How to) Call somatic mutations using GATK4 Mutect2
    Read Filters explained
    Read Filters Docs

    The default values set of Mutect2 are done after many tests on different sets of data. And for more information on what happens to the read you could look into the code base of the tool which you can find here.

Sign In or Register to comment.