Skip to main content

Table 6 Consensus when optimising for both datasets (excluding models trained on test data, excluding poor performing tools, no more than five tools, recall approx. 20%)

From: Improving CRISPR guide design with consensus approaches

  Wang   Doench
Consensus level n Accepted Precision Recall NPV Accepted Precision Recall NPV
1 811 0.724 0.803 0.598 1327 0.229 0.819 0.870
2 453 0.817 0.506 0.496 762 0.280 0.574 0.854
3 148 0.912 0.185 0.416 225 0.356 0.216 0.820
4 23 0.957 0.030 0.381 24 0.500 0.032 0.802
  1. Tools considered here: sgRNAScorer2, CHOPCHOP, PhytoCRISP-Ex, mm10db