28  March 2022
Jeffrey M. Kidd
University of Michigan


This directory contains structural variant for 1,879 samples processed by the Dog10K
Consortium.  This is a subset of the total samples as samples with uneven read depth or
excessive discordant read-pairs were removed. 

Calls were identified in each sample using Manta version 1.6.0 with default parameters.
Inversion were converted to event representation using the convertInversion.py utility. 
Raw calls were merged using svimmer and genotyped across all samples using GraphTyper2 
version 2.7.2 with default parameters. All analyiss was limited to the primary
assembled chromosomes.

For break-end (BND), insertion (INS), deletion (DEL) and duplication (DUP) calls, the 
‘AGGREGATED’ genotyping model was used.  For inversion (INV) candidates, the breakpoint 
model was used as reported by GraphTyper2. 

Variants were filtered using vcffilter with the following command: 

vcffilter -f "( SVTYPE = BND & SVMODEL = AGGREGATED & QD > 20 & ( ABHet > 0.30 | ABHet < 0 ) & ( AC / NUM_MERGED_SVS ) < 10 & PASS_AC > 0 & PASS_ratio > 0.1 ) | ( SVTYPE = DEL & SVMODEL = AGGREGATED & QD > 12 & ( ABHet > 0.30 | ABHet < 0 ) & ( AC / NUM_MERGED_SVS ) < 25 & PASS_AC > 0 & PASS_ratio > 0.1 ) | ( SVTYPE = DUP & SVMODEL = AGGREGATED & QD > 5 & PASS_AC > 0 & ( AC / NUM_MERGED_SVS ) < 25 ) | ( SVTYPE = INS & SVMODEL = AGGREGATED & PASS_AC > 0 & ( AC / NUM_MERGED_SVS ) < 25 & PASS_ratio > 0.1 & ( ABHet > 0.25 | ABHet < 0 ) & MaxAAS > 4 ) | ( SVTYPE = INV & PASS_AC > 0 & ( AC / NUM_MERGED_SVS ) < 25 & PASS_ratio > 0.1 & ( ABHet > 0.25 | ABHet < 0 ) & MaxAAS > 4 )" 

Variant Type    Pass GraphTyper2 Filters
BND 150
DEL 70799
DUP 3070
INS 67791
INV 3061


Note: a sample swap in the SV vcf genotypes for BMAL000003 and LAEK000002 was identified.  The posted filtered files have the correct genotype data.