You are viewing the site in preview mode

Skip to main content

Table 1 Performances (F1 scores in %) of SNP and indel predictions by NanoCaller, Medaka, Clair, Longshot, and DeepVariants on ONT and PacBio (CCS and CLR) datasets. This evaluation is based on v3.3.2 benchmark variants for HG001 and HG005-7, and v4.2.1 benchmark variants for the Ashkenazim trio (HG002, HG003, and HG004). Bonito and R10.3 refer to different versions of the HG002 ONT datasets

From: NanoCaller for accurate detection of SNPs and indels in difficult-to-map regions from long-read sequencing by haplotype-aware deep neural networks

Prediction Variant caller HG001 HG002 HG003 HG004 HG005 HG006 HG007 HX1 Bonito R10.3
SNPs on ONT data in high-confidence intervals NanoCaller ONT-HG001 98.58 98.35 98.99 98.97 98.10 98.23 97.81 98.45 99.33 98.28
NanoCaller ONT-HG002 98.63 98.66 99.09 99.11 98.38 98.43 98.06 98.60 99.34 98.44
Medaka 99.03 98.59 99.02 99.04 98.17 98.50 98.24 98.94 99.24 96.94
Clair 98.79 97.77 98.60 98.58 97.73 97.90 97.50 98.53 98.75 90.44
Longshot 98.78 98.03 97.88 97.90 98.34 98.53 98.51 98.59 98.59 98.18
Indels on ONT data in high-confidence intervals NanoCaller ONT-HG001 57.33 53.94 58.52 57.71 56.31 56.14 53.78 73.67 62.07 61.59
NanoCaller ONT-HG002 56.69 54.37 58.47 57.69 56.93 56.56 54.44 73.90 61.17 60.56
Medaka 48.67 48.10 53.59 50.19 55.89 52.49 51.83 81.13 51.03 53.09
Clair 49.72 47.64 52.06 51.20 52.58 51.90 50.63 80.59 50.11 44.80
Indels on ONT data in non-homopolymer regions NanoCaller ONT-HG001 87.65 82.28 87.93 87.93 81.92 85.70 83.41 59.47 86.12 84.43
NanoCaller ONT-HG002 87.19 82.80 87.93 88.04 82.60 86.10 83.92 59.17 85.76 83.51
Medaka 82.07 78.70 85.74 84.23 80.97 84.41 82.91 55.17 78.24 78.75
Clair 75.25 70.06 75.55 74.85 72.60 75.92 75.04 58.43 70.99 62.93
SNPs on PacBio CCS data in high-confidence intervals NanoCaller CCS-HG001 99.25 99.80 99.79 99.71       
NanoCaller CCS-HG002 99.17 99.80 99.79 99.75       
Clair 99.66 99.84 99.72 99.79       
Longshot 99.37 99.03 99.05 99.05       
DeepVariant 99.82 99.93 99.91 99.84       
Indels on PacBio CCS data in high-confidence intervals NanoCaller CCS-HG001 92.67 93.30 93.42 93.10       
NanoCaller CCS-HG002 93.13 94.10 94.34 93.97       
Clair 94.87 96.71 97.51 95.57       
DeepVariant 98.21 99.28 99.48 98.42       
SNPs on PacBio CLR data in high-confidence intervals NanoCaller CLR-HG002 94.42 98.75 94.41 93.41       
Clair 95.83 98.38 94.89 94.15       
Longshot 96.81 98.41 94.35 93.27       
Back to article page

Genome Biology

ISSN: 1474-760X

Contact us