[Assessment] Breakend distance in SV obs-frequency annotation

[mathiasbio]

Description

In our SV observation frequency annotation we're a bit inconsistent, and possibly quite too relaxed in our maximum break-end distance for considering a variant "the same". Currently it's 5000 for Swegen, and 10 000 for our Loqusdb annotations, that is from rare-disease and from somatic SVs. I want to make sure that this relaxed setting doesn't have too many risks of filtering out real somatic variants.

Suggested approach

As the goal behind this frequency annotation is to remove uninteresting variants and make interpretation easier, we can decrease this maximum breakend distance and understand how the number of variants changes. If we can reduce the maximum breakend distance to 1000 instead, and see that the number of variants do not increase beyond a manageable level, it should be fine to implement.

Criteria

No response

Origin

No response

Anything else?

No response

[mathiasbio]

I will check to see if the extra variants that pop up after increasing stringency of breakend distance should be filtered or not

[jemten]

Managed to catch @J35P312 and spontaneously he liked the settings as is. Will be interesting to hear what your test @mathiasbio

[mathiasbio]

Okay! That's nice to hear. That's enough reassurance for me to feel safe on putting this on the backburner for a while.