Deletions reported below the min-indel-size threshold

[ASLeonard]

Hi,

@lauradekker and I noticed there are variants that are shorter than --min-indel-size that are reported in both the discover and joint-call steps. The numbers are overall small and can easily be filtered by bcftools, but the behaviour is unexpected. It also appears to always be DELs.

The numbers below are based on 428,920 total SVs in the genotyped.sv.vcf.gz file for 166 cattle samples.

SVTYPE SVLEN COUNT
DEL 1 16
DEL 2 39
DEL 3 4
DEL 4 7
DEL 5 3
DEL 6 4
DEL 8 2
DEL 9 2
DEL 10 1
DEL 11 2
DEL 12 1
DEL 14 1
DEL 19 1
DEL 22 13
DEL 23 20
DEL 24 2
DEL 25 1
DEL 29 1

Alignments done with pbmm2 v1.17.0 and called with sawfish v2.2.1.

Best,
Alex