Figure 4.

Comparison of false positive predictions on the test dataset with 200 false arrays. In summary, all false arrays were reported as a true array by at least one existing tool. In contrast, CRISPRidentify predicts all 200 as true negatives. Even further, the majority (159 out of 200) of the arrays were reported as a true array by only one approach, validating the quality of the false data set all these 159 arrays where reported as false arrays by two existing tools. For more details about the datasets see corresponding statistics in the Supplementary Table S4 Section 10. Corresponding distribution plots are presented in Supplementary Section 19.