Figure 2| Computational model predicting guide RNA activities
(a) (b) The schematic of datasets and algorithm used
(c) ROC curves comparing the predictive power of our model and deepCpf1.
(d) A quantization of (c) and the bar shows the mean ±s.d. between predicted and measured guide activity scores (n=10)
(e) Comparison of the different models. A smaller scale library of 6,265 guide RNAs was processed by each model, and the top 1,000 scored ones were chosen as an efficient group, in comparison to bottom 1,000 ones as an inefficient group. Their corresponding experimental activity scores were compared.
(f) The former library was re-chosen into high score and low score group by measured guide RNA activities (1,000 members from 6,265).