A few instances still existed where 'get me the cohort part of the config' wasn't explicitly called with a dataset parameter, so it defaulted to seqr
Seqr is not a valid cohort, so weird error messages were popping up e.g.
assert COHORT_SEQ_CONFIG, f'{dataset} - {seq_type} is not represented in config'
AssertionError: seqr - exome is not represented in config
Proposed Changes
checks that a dataset override is passed to all possible instances of the config getter method(s)
fixes VEP memory setting (requested 16GB, hail default VM is highmem @ 6.5GB RAM/core. To solve the RAM/core ratio at 6.5GB/core, 16GB is only solved by 4 cores, giving 26GB, which is way too much. See https://batch.hail.populationgenomics.org.au/batches/429539/jobs/1)
Fixes
seqr
Proposed Changes