purity_estimate mode: validator incorrectly requires normal REDUX BAM after 2ba879d1

[gavin-peng]

Description of the bug

Hi @luan-username — this looks related to your recent commits on the purity estimate WGTS mode.

Commit 2ba879d updated AMBER_PROFILING so that in purity_estimate mode, AMBER is skipped when no normal REDUX BAM is provided (and WISP runs without AMBER_LOH).
However, checkLongitudinalSampleInputs in SampleSheet.groovy still throws error when we intentionally omitting normal BAM entirely :

if (longitudinal && has_amber_dir && !has_normal_dna_bam) {
throw ...
}

This means the validator blocks a configuration the subworkflow now explicitly handles gracefully.

Suggested fix: rather than requiring a normal REDUX BAM, only block a raw (non-REDUX) normal BAM — which genuinely cannot be used for longitudinal AMBER — while
allowing the no-normal path through:

def has_normal_raw_bam = Inputs.hasNormalDnaBam(meta)
if (longitudinal && has_amber_dir && has_normal_raw_bam) {
throw new IllegalStateException(
"group_id(${meta.group_id}) - In purity estimate mode, the primary normal DNA BAM " +
"must be REDUX-processed (bam_redux filetype); a raw BAM is not accepted"
)
}

This preserves the April 28 (c05ebe7) enforcement that raw BAMs are rejected, while allowing the no-normal path introduced April 29.

Command used and terminal output

Relevant files

No response

System information

No response

[luan-n-nguyen]

Hi Gavin,

This is intended - the amber_dir and normal bam_redux must go together. If you want to omit the normal bam_redux, you must also omit the amber_dir to avoid the error.

Your suggested change means that providing amber_dir without the normal REDUX BAM results in AMBER silently being skipped (due to this AMBER_PROFILING check: def runnable_purity_estimate = purity_estimate_mode && normal_bam). I agree that the check in SampleSheet.checkLongitudinalSampleInputs is redudant to the AMBER_PROFILING check, but the SampleSheet check is implemented that way so that the error is explicit.

Luan

[gavin-peng]

Thanks Luan,
I undersdtand now. then we will just make adjustment on our end (the wdl wrapper) to skip amber_dir and normal REDUX_BAM together.
Thanks for your very fast responce!