medaka variant produces unsorted vcf, crashing whatshap about medaka

[cjw85]

Hi @wdecoster,

Can you provide a minimal example (bam) that reproduces this? Our current pre-release may fix this; we've worked quite a bit on how chunks of data are handled. The pre-release is available on pypi: pip install medaka==0.7.0a1. If you want to test this, note that some command line options may have changed.

[wdecoster]

I encountered this issue for a 20x human genome coverage bam, but not for two larger bams. The bam is 80Gb, which is perhaps not a 'minimal' example. I'll have a look at the pre-release.

[cjw85]

Wouter can you confirm the v0.7.x releases resolve this issue for you?

[wdecoster]

I'm now on v0.7.0-alpha.1, and running in parallel in separate windows over the genome (I should maybe look into using the GPUs of our PromethION...).
I haven't encountered the error anymore, but I also haven't exactly reproduced what caused in the first place, so I cannot confirm conclusively that it is resolved.

[cjw85]

Thanks, I believe the issue is resolved: I have run successfully through a chrom1 dataset.

Using a GPU is definitely preferred if available; a V100 will perform the consensus steps at around 10Mb per minute.