GenotypeGVCFs --only-output-calls-starting-in-intervals combined with -L . What does that mean?

In GATK4,

the documentation for GenotypeGVCFs argument --only-output-calls-starting-in-intervals mentions:

Restrict variant output to sites that start within provided intervals
This option can only be activated if intervals are specified.

I am using GenotypeGVCFs repeatedly on every non-overlapping interval -L foo-bar to cover the genome. Does that documentation mean that for calls which "straddle" two consecutive intervals A and B, the calls will be present in the tail end of A.vcf, and not in B.vcf? Is that the correct way to interpret the manual?

Similarly, if I omitted --only-output-calls-starting-in-intervals, would that mean that the calls are duplicated VCFs if they land on the edge of two intervals?

Comments

  • SheilaSheila Broad InstituteMember, Broadie, Moderator admin

    @init_js
    Hi,

    Does that documentation mean that for calls which "straddle" two consecutive intervals A and B, the calls will be present in the tail end of A.vcf, and not in B.vcf? Is that the correct way to interpret the manual?

    Yes, that sounds correct.

    Similarly, if I omitted --only-output-calls-starting-in-intervals, would that mean that the calls are duplicated VCFs if they land on the edge of two intervals?

    That sounds right. Perhaps this ticket will help as well.

    -Sheila

  • blueskypyblueskypy Member ✭✭
    edited February 25

    Sounds to me -L should always go with --only-output-calls-starting-in-intervals in any GATK function producing variant calling, is it correct?

  • bhanuGandhambhanuGandham Cambridge MAMember, Administrator, Broadie, Moderator admin

    @blueskypy

    Sounds to me -L should always go with --only-output-calls-starting-in-intervals in any GATK function producing variant calling, is it correct?

    Yes that is correct.

Sign In or Register to comment.