Fig. 3: Analysis of differences in tracrRNA activity.

a SHAP feature importance for the 10 most important features for predicting the difference between sgRNAs screened with the Hsu versus Chen tracrRNA. Each point represents one sgRNA from the training set. Descriptions of model features can be found in Supplementary Data 3. b Box plot of the difference in z-score log-fold changes between sgRNAs screened with the Hsu versus Chen tracrRNA as a function of G/T presence in positions 17–20 of each spacer sequence. The ‘~’ symbol indicates the nucleotide is not present in this range. Box minima and maxima represent the 25th and 75th percentile, whisker minima and maxima represent the 25th and 75th percentile minus and plus 1.5× the interquartile range, respectively, and the box centers represent the 50th percentile of data. c Box plot of the difference in z-score log-fold changes between sgRNAs screened with the Hsu versus Chen tracrRNA as a function of G/T abundance in positions 17-20 of each spacer sequence. Box minima and maxima represent the 25th and 75th percentile, whisker minima and maxima represent the 25th and 75th percentile minus and plus 1.5× the interquartile range respectively, and the box centers represent the 50th percentile of data. d Same as (b) but for the Hsu and DeWeirdt tracrRNAs. e Same as (c) but for the Hsu and DeWeirdt tracrRNAs. f Data from Mimitou et al. Comparison of sgRNA relative abundance when read out via gDNA or direct sequencing. Spacers are binned by G/T abundance in positions 17–20. Source data are provided as a Source Data file.