search

arq5x / lumpy-sv

lumpy: a general probabilistic framework for structural variant discovery
MIT License
309 stars 118 forks source link

Multiple calls in same region #251

Closed renzobalb closed 6 years ago

renzobalb commented 6 years ago

Hi Ryan,

I had a question regarding lumpy calls -

In my dataset, I managed to find a bunch of separate calls that overlap considerably.

For example (this is a subset of one sample, on the same chromosome):

chr start end length QUAL SVType chr 11014836 11070622 55786 25.33 DUP chr 11015613 11070363 54750 119.60 DEL chr 11016535 11072796 56261 202.60 DUP chr 11017350 11072304 54954 47.58 DEL chr 11025022 11052632 27610 35.51 DUP chr 11026167 11052463 26296 203.70 DEL chr 11028884 11132399 103515 96.91 DUP chr 11029840 11061321 31481 27.17 DUP

The vast majority of the variants also have the same genotype called using SVTyper.

I was wondering as to what causes this, how accurate these are, and the best practice for dealing with this situation (e.g. do I merge these variants into one/ by type etc.)?

Thanks, Renzo

ryanlayer commented 6 years ago

Try visualizing that region with samplot. Could be a noisy region or complex SV.

https://github.com/ryanlayer/samplot

On Jun 19, 2018, at 1:13 AM, rbalb notifications@github.com wrote:

Hi Ryan,

I had a question regarding lumpy calls -

In my dataset, I managed to find a bunch of separate calls that overlap considerably.

For example (this is a subset of one sample, on the same chromosome):

chr start end length QUAL SVType chr 11014836 11070622 55786 25.33 DUP chr 11015613 11070363 54750 119.60 DEL chr 11016535 11072796 56261 202.60 DUP chr 11017350 11072304 54954 47.58 DEL chr 11025022 11052632 27610 35.51 DUP chr 11026167 11052463 26296 203.70 DEL chr 11028884 11132399 103515 96.91 DUP chr 11029840 11061321 31481 27.17 DUP

The vast majority of the variants also have the same genotype called using SVTyper.

I was wondering as to what causes this, how accurate these are, and the best practice for dealing with this situation (e.g. do I merge these variants into one/ by type etc.)?

Thanks, Renzo

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

renzobalb commented 6 years ago

Great, thanks - looks to be a complex SV.