search

arq5x / lumpy-sv

lumpy: a general probabilistic framework for structural variant discovery
MIT License
307 stars 119 forks source link

PE=0 for all SVs derived from 10X Long Ranger bams. #338

Open andrewSharo opened 4 years ago

andrewSharo commented 4 years ago

Hi Ryan,

I'm running lumpy v0.2.13 because I need to compare to some earlier analysis. When I run lumpyexpress, it outputs a vcf where all records have 0 PE support. My sample.splitters.bam and sample.discordants.bam files look normal (9 Gb and 14Gb) and are in order according to the check_sorting.py script.

The bams are from 10X if that helps, so not aligned with bwa. Are there certain bam tags that lumpy requires for PE support? The bams look normal to me.

Best, Andrew

ryanlayer commented 4 years ago

We have never really tested anything but bwa aligned bams in a very long time. We did novoalign at first but dropped it once bwa mem came along. That is all to say that this could be a bwa issue. We do filter on certain flags and if those flags are set differently then those reads may get dropped.

Are you sure the file was aligned as pairs?

On Jun 17, 2020, at 4:26 PM, andrewSharo notifications@github.com wrote:

 Hi Ryan,

I'm running lumpy v0.2.13 because I need to compare to some earlier analysis. When I run lumpyexpress, it outputs a vcf where all records have 0 PE support. My sample.splitters.bam and sample.discordants.bam files look normal (9 Gb and 14Gb) and are in order according to the check_sorting.py script.

The bams are from 10X if that helps, so not aligned with bwa. Are there certain bam tags that lumpy requires for PE support? The bams look normal to me.

Best, Andrew

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

andrewSharo commented 4 years ago

Hi Ryan, Normally, that would make sense to me, However I've also run this same bam through smoove (v0.2.5) and it finds plenty of PE evidence (although not a lot of SR evidence). Is there a big difference between the way smoove and lumpyexpress read bams? Best, Andrew

ryanlayer commented 4 years ago

Ah. Yes. The bam filtering is much improved. Use smoove. We are ending support for lumpy express.

On Jun 18, 2020, at 12:39 AM, andrewSharo notifications@github.com wrote:

 Hi Ryan, Normally, that would make sense to me, However I've also run this same bam through smoove (v0.2.5) and it finds plenty of PE evidence (although not a lot of SR evidence). Is there a big difference between the way smoove and lumpyexpress read bams? Best, Andrew

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.