Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.835 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I70; Method DRAGENv4.2; Interval 100-499; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.948 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I72; Method DRAGENv4.2; Interval 1000-4999; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.909 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I74; Method DRAGENv4.2; Interval 10000-19999; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.767 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I69; Method DRAGENv4.2; Interval 50-99; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.888 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I71; Method DRAGENv4.2; Interval 500-999; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.96 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I73; Method DRAGENv4.2; Interval 5000-9999; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.828 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I76; Method DRAGENv4.2; Interval ALL; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.75 f1-scorefraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, I75; Method DRAGENv4.2; Interval >20000; column F1Score Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.897 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G70; Method DRAGENv4.2; Interval 100-499; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.992 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G72; Method DRAGENv4.2; Interval 1000-4999; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 1 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G74; Method DRAGENv4.2; Interval 10000-19999; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.855 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G69; Method DRAGENv4.2; Interval 50-99; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.959 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G71; Method DRAGENv4.2; Interval 500-999; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 1 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G73; Method DRAGENv4.2; Interval 5000-9999; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.899 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G76; Method DRAGENv4.2; Interval ALL; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.818 precisionfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, G75; Method DRAGENv4.2; Interval >20000; column Precision Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.78 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H70; Method DRAGENv4.2; Interval 100-499; column Recall Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.907 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H72; Method DRAGENv4.2; Interval 1000-4999; column Recall Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.833 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H74; Method DRAGENv4.2; Interval 10000-19999; column Recall Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.695 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H69; Method DRAGENv4.2; Interval 50-99; column Recall Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.827 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H71; Method DRAGENv4.2; Interval 500-999; column Recall Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.923 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H73; Method DRAGENv4.2; Interval 5000-9999; column Recall Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.768 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H76; Method DRAGENv4.2; Interval ALL; column Recall Configuration: DRAGENv4.2 (Nardone et al.) Protocol: HG002 short-read deletion calling by length bin, witty.er (Nardone et al. Table S1) Dataset: HG002 GIAB SV v0.6 Tier 1 deletions lifted to GRCh38 (5,414) 0.692 recallfraction · higher
Uncertainty: Not reported by the source
Coverage: Not reported scored / Not reported eligible
Independent external evaluation · Source checked Evidence concern: excluded from comparisons Results 3.1 prose gives inGAP in the 5000-9999 bp bin as F1 97%, precision 92%, recall 94%. Table S1 rows for inGAP 5000-9999 print F1 0.9466, precision 0.9714, recall 0.9231. The prose values are not consistent with each other (F1 cannot exceed both precision and recall) or with the table. Table values are recorded; prose values are not.A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies · Results 3.1 paragraph 2 versus TableS1.xlsx Method inGAP, Interval 5000-9999
Short-read input depth and alignment for Table S1 are stated inconsistently: Methods 2.2 says the 65x reads were subsampled to 25x and aligned with bwa-mem2 for preliminary evaluations; Results 3.1 says the HG002 data were previously aligned using the DRAGEN pipeline; the Data Availability Statement cites 30x Illumina HG002 2x250 reads. Which depth and aligner produced Table S1 is not resolved, so these results should not be compared automatically with other HG002 tables.Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · Methods 2.2 paragraph 1; Results 3.1 paragraph 1; Data Availability Statement
Methods, coverage and source DRAGENv4.2 on HG002 deletions (Nardone et al.)
cnv-20261009-protocol-nardone2025-hg002-del-wittyer
Aggregation: Not reported
A Hitchhiker Guide to Structural Variant Calling: A Comprehensive Benchmark Through Different Sequencing Technologies ; Nardone et al. 2025, Table S1 (ten short-read SV callers on HG002) · TableS1.xlsx, H75; Method DRAGENv4.2; Interval >20000; column Recall