Skip to main content

Table 11 Indel results for HG003 chr21-22, Hybrid

From: HELLO: improved neural network architectures and methodologies for small variant calling

  DeepVariant HELLO
Precision Recall Errors Precision Recall Errors
20 × 15x 0.986837 0.984444 401 0.991683 0.989605 261
20 × 30x 0.988237 0.989823 308 0.995143 0.993676 156
20 × 60x 0.989732 0.991568 263 0.997114 0.995929 97
30 × 15x 0.992761 0.991059 226 0.996052 0.993458 146
30 × 30x 0.992854 0.993458 192 0.996902 0.995639 104
30 × 60x 0.992244 0.993821 196 0.997958 0.996802 73
40 × 15x 0.995501 0.993312 156 0.996622 0.995202 114
40 × 30x 0.994741 0.994839 146 0.997608 0.996874 77
40 × 60x 0.992944 0.994839 172 0.997258 0.997165 78
50 × 15x 0.995572 0.993967 146 0.997323 0.995202 104
50 × 30x 0.995649 0.995348 126 0.997257 0.996656 85
50 × 60x 0.993425 0.995348 158 0.997890 0.997383 66
  1. Values in bold indicate the best operating point for each tool for the corresponding metric