Tim O'Donnell

internal server error creating new run

1

Sorry this error isn't more informative. In our hammerlab cycledash instance, when I try to Add a Run with these files, I get an "internal server error": /user/ahujaa01/brody-lymphoma/brody_442/brody_442-bcell-tumor-bwa-mem-gap11-gep4-dedup-indelrealigned-bqsr.bam /user/ahujaa01/brody-lymphoma/brody_442/brody_442-tcell-normal-bwa-mem-gap11-gep4-dedup-indelrealigned-bqsr.bam /user/ahujaa01/brody-lymphoma/brody_442/brody_442-bcell-bwa-mem-gap11-gep4-dedup-indelrealigned-bqsr-Mutect-default-merged.vcf...

feature request: have a way to show gene, intron, exon annotation tracks

3

Also, a way to jump to a gene by typing it in somewhere would be really useful. My prompt for this is I am scrolling through TP53 in Cycledash /...

Just noting this in case it matters later, regarding the `dbsnp_alt_url` and `dbsnp_broad_url` defined here to be passed to mutect: https://github.com/hammerlab/biokepi/blob/master/src/lib/run_environment.ml#L656-L657 - the `dbsnp_alt_url` is a dead link: ftp://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606/VCF/v4.0/00-All.vcf.gz -...

I'm not quite sure where to file this ticket, could also go in internal-tools perhaps. I'm finding it difficult to get my head around biokepi. Lots of this is me...

For some analyses we're doing, I'd like to be able to run variant callers on randomly downsampled data. An example analyses would be: run MuTect on subsamples of my BAM...

Fix typo in field descriptions

low GPU utilization on multi-GPU inference due to synchronization

2

I mentioned this to @eric-czech in my org and he offered to fix this, putting details for Eric here: Boltz1 does multi-GPU inference using a pytorch lightning Trainer. Each prediction...

boltz predict should fail (return nonzero) when it runs out of memory

2

It would be helpful if `boltz predict` would return a nonzero status when it fails with errors like: ``` LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0] Predicting DataLoader 0: 0%| | 0/1...

A few small things: 1. Fixed an issue where affinity predictions were being re-used (not rerun) even when --override is passed 1. Added support for passing --devices=0 to use all...