Maximum alt alleles guardrail [VS-1530] #9033
Conversation
| @@ -135,9 +135,9 @@ Once the VAT table is created and a tsv is exported, the AoU research workbench | |||
| - Specify the same `call_set_identifier`, `dataset_name`, `project_id`, `extract_table_prefix`, and `interval_list` that were used in the `GvsPrepareRangesCallset` run documented above. | |||
| - Specify the `interval_weights_bed` appropriate for the PGEN / VCF extraction run you are performing. `gs://gvs_quickstart_storage/weights/gvs_full_vet_weights_1kb_padded_orig.bed` is the interval weights BED used for Quickstart. | |||
| - For both `GvsExtractCallset` and `GvsExtractCallsetPgenMerged`, select the workflow option "Retry with more memory" and choose a "Memory retry factor" of 1.5 | |||
| - `GvsExtractCallset` currently defaults to 1000 alt alleles maximum, which means that any sites having more than that number of alt alleles will be dropped. Make sure to specify the appropriate `maximum_alternate_alleles` value (currently 100). | |||
There was a problem hiding this comment.
I'm confused. "currently 100?"
There was a problem hiding this comment.
This is carried over from line 140. The value of max alt alleles we were told to use for Echo was 100, but in my testing we seem to be able to go to 1000 at least without WeightedSplitIntervals breaking down. I don't know if variants with > 100 alt alleles are actually useful though.
There was a problem hiding this comment.
for posterity's sake, I'll clarify that WeightedSplitIntervals does not actually check alleles at all. I know we discussed this and everyone involves knows now, but juuuuust in case someone goes back to this PR in the future I want it to be stated here too lol
There was a problem hiding this comment.
I do agree with George I think that there are two values for currently in this sentence.
currently defaults to 1000
and
(currently 100).
and to me that is the confusing part
mcovarr
changed the title
Extract and WeightedSplitIntervals don't deal well with arbitrarily large numbers of alt alleles, so add an (overridable) guardrail of 1000 alt alleles max if the user neglects to specify the 100 alt alleles max recommended in our documentation for AoU callsets.