Skip to content

Multi-sample phasing in glimpse2_phase_cram #240

@tefirman

Description

@tefirman

Bug Description

Which module or pipeline has the issue?
ww-glimpse2, ww-imputation

What version are you using?
Latest from main

What went wrong?
The glimpse2_phase_cram task only accepts a single BAM/CRAM file via --bam-file, but GLIMPSE2_phase supports joint multi-sample phasing via --bam-list. The ww-imputation pipeline works around this by scattering over individual samples and phasing each one independently, which doesn't match the recommended usage pattern from the Palantir/Broad GLIMPSE2 workflow.

What did you expect to happen?
glimpse2_phase_cram should accept an array of BAM/CRAM files and use --bam-list to phase all samples jointly in a single call. The ww-imputation pipeline should pass all input CRAMs into each phase call rather than scattering per-sample, producing a single multi-sample output file.

Additional Context

Anything else that might be helpful?

  • GLIMPSE2_phase docs describe --bam-file (single) vs --bam-list (multiple) as mutually exclusive input modes
  • The fix involves a breaking change to both the ww-glimpse2 task signature (FileArray[File]) and the ww-imputation pipeline inputs (removing ImputationSample struct in favor of flat arrays)
  • Both modules are at prototype status so breaking changes are expected

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions