Skip to main content
. 2022 Jan 12;11:e73577. doi: 10.7554/eLife.73577

Table 2. Site-specific and sample-specific filters used by the different groups to detect de novo mutations (DNMs) in Heineken (difference in the other steps of the pipeline in Table 2—source data 1).

Table 2—source data 1. Details on the methodology and filtering criteria applied by the five different pipelines to estimate the mutation rate on the common pedigree.
Research group Candidate DNMs Site-specific filters Sample-specific filters Additional filters
CV 18 GATK Best Practices
hard filter criteria
0.5 × dpind < DP < 2 ×
dpind
GQ > 40
AD > 0
0.25 < AB < 0.75
RW 22 QD < 2.0
MQ < 40.0
FS > 60.0
SOR > 3.0
MQRankSum < –12.5
ReadPosRankSum < –8.0
20 < DP < 80
GQ > 20
AD > 0
0.35 < AB
Alternative allele on both strands
TT 27 Remove variants in recent repeats or in homopolymers of AAAAAAAAAA or TTTTTTTTTT DP > 10
GQ > 20
AD > 0
0.25 < AB
Overlap three different variant callers

Filter on LCR
LB 28 QD < 2.0
FS > 20.0
MQ < 40.0
MQRankSum < –2.0
MQRankSum > 4.0
ReadPosRankSum < –3.0
ReadPosRankSum > 3.0
SOR > 3.0
0.5 × dpind < DP < 2 ×
dpind
GQ > 60
AD none
0.3 < AB < 0.7
Manual curation (six candidates removed)
SB 32 FS > 30.0
MQRankSum < –10
MQRankSum > 10
ReadPosRankSum < –2.5
ReadPosRankSum > 2.5
BaseQRankSum < –13
BaseQRankSum > 13
10 < DP < 2× dpind
GQ > 55
AD > 0
0.3 < AB
Alternative allele in both strands. lowQ AD2 > 1

LB: Lucie Bergeron; SB: Søren Besenbacher; CV: Cyril Versoza; TT: Tychele Turner; RW: Richard Wang.